Road surface feature recognition method based on electronic mechanical brake vehicle

Abstract

The method comprises the following steps: judging whether each wheel clamping force sensor works normally, if so, directly obtaining a brake clamping force as a final brake clamping force, if not, obtaining a first brake clamping force through a motor model, and if not, obtaining a second brake clamping force through a motor model; second brake clamping force is obtained through the brake friction plate model, and fusion is conducted through the fuzzy logic weight algorithm to obtain final brake clamping force; judging whether the wheel has a locking risk or not, if not, calculating the current ground longitudinal force of the wheel according to a kinetic model of the wheel, if so, triggering longitudinal force locking logic, and acquiring the current locking ground longitudinal force as the current ground longitudinal force of the wheel; calculating the current utilization adhesion coefficient of each wheel in combination with the vertical force; and determining the road surface type of the vehicle by using a fuzzy rule based on the adhesion coefficient of each wheel. According to the invention, the problem of low identification precision is solved.

Description

Technical Field

[0001] This invention relates to the field of automotive electromechanical braking technology, and more specifically to a method for road surface feature recognition of vehicles based on electromechanical brakes. Background Technology

[0002] Vehicle driving safety is closely related to road surface characteristics. Accurately identifying road surface types, such as uniform roads, split roads, and connecting roads, is a core prerequisite for vehicle stability control and braking system optimization. Traditional road surface recognition methods mostly rely on additional sensors such as onboard cameras and LiDAR, which suffer from high costs, significant susceptibility to ambient lighting, and slow response times.

[0003] Electromechanical brakes (EMBs), as a new braking technology, can directly acquire key parameters such as clamping force and motor torque, providing a new data source for road feature recognition.

[0004] However, existing EMB-based identification methods rely on sensors for clamping force measurement, and lack reliable redundancy estimation schemes when sensors fail, resulting in low identification accuracy when the methods are applied.

[0005] Therefore, there is an urgent need for a road surface feature recognition method based on vehicles with electromechanical brakes to solve the problem that traditional EMB-based road surface feature recognition methods rely solely on sensors to measure clamping force, resulting in low recognition accuracy. Summary of the Invention

[0006] This invention addresses the shortcomings of existing technologies by providing a road surface feature recognition method for vehicles with electromechanical brakes. This method solves the problem that traditional EMB-based road surface feature recognition methods rely solely on sensors to measure clamping force, resulting in low recognition accuracy.

[0007] To achieve the above objectives, the present invention adopts the following technical solution: Determine whether the clamping force sensors of the electromechanical brakes of each wheel are working properly. If the clamping force sensors are working properly, the brake clamping force is directly obtained as the final brake clamping force. If the clamping force sensors fail, the first brake clamping force is obtained through the motor model, the second brake clamping force is obtained through the brake friction pad model, and the final brake clamping force is obtained by fusing them through the fuzzy logic weight algorithm. After obtaining the final brake clamping force, it is determined whether the wheel is at risk of locking up. If there is no risk of locking up, the current ground longitudinal force of the wheel is calculated based on the wheel's dynamic model. If there is a risk of locking up, the longitudinal force locking logic is triggered, and the current locked ground longitudinal force is obtained. The obtained locked ground longitudinal force is used as the current ground longitudinal force of the wheel. The current coefficient of adhesion for each wheel is calculated based on the longitudinal ground force and the final brake clamping force of each wheel. Based on the adhesion coefficient of each wheel, fuzzy rules are used to determine the type of road surface the vehicle is on.

[0008] To optimize the above technical solution, the specific measures also include: Furthermore, obtaining the clamping force of the first brake through the motor model includes the following steps: A motor model based on the torque of a permanent magnet synchronous motor is constructed, and the clamping force of the first brake is calculated. : ; In the formula, This represents the q-axis current of the permanent magnet synchronous motor. For extreme logarithms, It is a permanent magnet flux linkage. For the moment of inertia of the wheel, The damping coefficient is... For mechanical angular velocity, Angular acceleration, The gear ratio of the reduction mechanism. This refers to the lead of the leadscrew.

[0009] Furthermore, obtaining the second brake clamping force through the brake friction pad model includes the following steps: A brake friction pad model based on friction pad deformation is constructed, and the total angular displacement of the motor during braking is decomposed to obtain the angular displacement that eliminates mechanical backlash. Angular displacement of friction plate undergoing elastic deformation And calculate the second brake clamping force generated by the deformation of the friction pad. for: ; In the formula, , , and This is the nonlinear stiffness coefficient.

[0010] Furthermore, the step of obtaining the final brake clamping force through fuzzy logic weighting algorithm fusion includes the following steps: The final brake clamping force is obtained by fusing the results using a fuzzy logic weighting algorithm. for: In the formula, As a weighting factor, Take [0,1].

[0011] Furthermore, the weighting factor The determination includes the following steps: The preset maximum clamping force is Define clamping force residual The maximum residual value is limited to ; Preset a fixed cycle and define the clamping force of the previous cycle. The three membership functions , and , and These represent the clamping force in the previous cycle. The clamping force residual is defined as low, medium, and high. The three membership functions , and , and These represent the clamping force residuals, respectively. Three levels: small, medium, and large; defining weighting factors. The three membership functions , and , , and They represent the weighting factors respectively. The three degrees of displacement dominance, equilibrium, and torque dominance; Established Weighting factors of the rules The fuzzy rule base, then for the first The rules are used to calculate the activation strength. : ; In the formula, This indicates that the smaller value on either side of the sign is taken as the result of the operation. Indicates the first The clamping force in the previous cycle of the rule. membership function, The range of values ​​is and , Indicates the first Clamping force residual in the rule of the clause membership function, The range of values ​​is and ; Aggregate the output fuzzy set and use the centroid method for defuzzification, then adjust the weight factors. The universe of discourse [0,1] is divided into 100 discrete points, and a weighted sum is calculated. The weighting factors are determined based on the calculation results. : ; ; In the formula, Indicates to The result of this rule is taken as a "fuzzy OR". Indicates according to the first Weighting factors determined by the rules The corresponding membership function output, The range of values ​​is and In the formula, For the first A discrete point.

[0012] Furthermore, determining whether the wheel is at risk of locking up includes the following steps: Introduce a death trend index It is used to assess the changing trend of wheel angular deceleration; when If the wheel is locked, it is determined that there is a risk of wheel lock-up; otherwise, there is no risk of wheel lock-up. This represents the rate of change of wheel angular deceleration.

[0013] Furthermore, if there is no risk of wheel lock-up, the longitudinal force on the ground of the wheel is calculated based on the wheel's dynamic model, including the following steps: Based on the dynamic model of the wheel, the longitudinal force on the ground is obtained. Clamping force of the final brake The relationship is: ; In the formula, The coefficient of friction of the brake disc friction surface. The effective braking radius of the brake disc. For the moment of inertia of the wheel, The angular deceleration of the wheel, The longitudinal force exerted on the wheel by the ground. Let be the rolling radius of the wheel.

[0014] Furthermore, if there is a risk of seizure, the longitudinal force locking logic is triggered, and the current locked ground longitudinal force is obtained, including the following steps: When a risk of wheel lockup is detected, the longitudinal force locking logic is triggered, and the initial wheel speed at the current moment is recorded. And continuously monitor wheel speed When the wheel speed decreases to a preset threshold wheel speed At this point, it is considered that the wheel has entered the deep slip phase, at which time the longitudinal force on the ground is... As it approaches the road surface adhesion limit, the final brake clamping force is then determined. Calculate the longitudinal force locking the ground. for: ; In the formula, The wheel speed is The angular deceleration of the wheel at that time, The coefficient of friction of the brake disc friction surface. The effective braking radius of the brake disc. For the moment of inertia of the wheel, Let be the rolling radius of the wheel.

[0015] Furthermore, the step of calculating the current coefficient of adhesion for each wheel based on the longitudinal ground force and the final brake clamping force includes the following steps: Based on the linear two-degree-of-freedom model of the vehicle, the height of the center of mass is set to... The wheelbase between the front and rear axles is set to , In the formula, This is the distance from the vehicle's center of gravity to the front axle. This is the distance from the vehicle's center of gravity to the rear axle. Based on the longitudinal force on the ground of each wheel That is, the longitudinal force on the left front of the ground. Longitudinal force on the right front ground Longitudinal force on the left rear ground and right rear longitudinal force And the final brake clamping force of each wheel. That is, the clamping force of the left front final brake. Right front final brake clamping force Left rear final brake clamping force and right rear final brake clamping force The adhesion coefficients of the left front were obtained respectively. Right front utilizes adhesion coefficient Left and rear use adhesion coefficient And right rear use of adhesion coefficient : ; In the formula, The coefficient of friction of the brake disc friction surface. Let be the vehicle mass, 'a' be the vehicle braking deceleration, and 'g' be the acceleration due to gravity. The effective braking radius of the brake disc. The rolling radius of the wheel, For the moment of inertia of the wheel, The angular deceleration of the left front wheel, The angular deceleration of the right front wheel. The angular deceleration of the left rear wheel, This is the angular deceleration of the right rear wheel.

[0016] Furthermore, the method of determining the road surface type of the vehicle based on the adhesion coefficient of each wheel using fuzzy rules to provide road surface condition information for the vehicle stability control system includes the following steps: determining the road surface type based on the adhesion coefficient of the left front wheel... Right front utilizes adhesion coefficient Left and rear use adhesion coefficient And right rear use of adhesion coefficient Calculate three core features: ; ; ; In the formula, The overall average of the adhesion coefficients is used for all four wheels. This represents the difference in the coefficient of adhesion between the left and right wheels. The difference in the coefficient of adhesion between the front and rear axle wheels; Next to , and Normalization is performed to obtain the corresponding result. , and ; Define the difference variables The fuzzy set is {small( ),middle( ),big( )}, and determine three levels of difference variables. The membership function, and and All used difference variables Membership function; defining the mean variable The fuzzy set is {low( ),middle( ),high( )}, and determine the mean variable for three degrees. Membership function; define road surface features The fuzzy set is {uniform low ( ), all in one middle ( ), uniformly high ( ), double ( ), docking ( ),mix( )}, and identified six levels of road surface characteristics. The membership function, and at the same time, the road surface features The elements in the fuzzy set correspond to the integers 1, 2, 3, 4, 5, and 6, respectively. Based on the road adhesion coefficient characteristics of each wheel, formulate The first fuzzy rule is determined; using a fuzzy inference mechanism, inference is performed based on the fuzzified input and fuzzy rules, and for each fuzzy rule, a product method is used to calculate the first fuzzy rule. The activation intensity corresponding to each fuzzy rule: ; In the formula, and They are the first The corresponding fuzzy rules , and membership function, The range of values ​​is and , The range of values ​​is and , The range of values ​​is and ; Based on activation intensity Aggregate the output fuzzy set to include road surface features Divide the range [1,7] into 100 discrete points and calculate the weighted sum; the resulting... The system uses a floor function to round down and outputs the final result. The final result is then mapped to road surface features. The integers corresponding to the elements in the fuzzy set are used to determine the current road surface conditions: ; ; In the formula, For the first The output road surface features corresponding to each rule membership function, The range of values ​​is , , , , and ; For the first Road surface features The discrete points.

[0017] The beneficial effects of this invention are: This invention employs a dual-mode approach of "direct measurement and dual-model redundancy estimation" to obtain the brake disc clamping force. When the sensor fails, the force can be estimated by combining the motor torque model, the friction pad deformation model, and the fuzzy logic weight algorithm, thus avoiding identification interruptions caused by reliance on a single sensor and ensuring the continuity of core data.

[0018] This invention introduces a wheel lock-up tendency index and longitudinal force locking logic. When a wheel is at risk of locking up, it can lock the longitudinal force value that is close to the road surface adhesion limit. This can solve the problems of the traditional method where the accuracy of ground longitudinal force calculation drops sharply when the wheel is locked up and the estimation using the adhesion coefficient has large deviation.

[0019] This invention eliminates the need for additional onboard cameras, lidar, or other sensors. It directly relies on parameters such as clamping force, motor torque, and wheel speed that can be obtained from the electromechanical brake itself to achieve identification. This reduces hardware costs and avoids the problem of external sensors being affected by environmental factors such as light and weather.

[0020] This invention achieves multi-dimensional data processing through fuzzy control logic. From clamping force fusion and calculation using the adhesion coefficient to road type judgment based on "overall mean, left-right difference, and front-back difference", all output results through efficient algorithm reasoning and defuzzification operations. It can quickly and accurately identify uniform road surfaces, split road surfaces, connected road surfaces, and mixed road surfaces, providing timely and reliable road condition information for vehicle stability control. Attached Figure Description

[0021] Figure 1 This is a schematic flowchart of a road surface feature recognition method for vehicles based on electromechanical brakes proposed in this invention. Detailed Implementation

[0022] The invention will now be described in further detail with reference to the accompanying drawings.

[0023] As attached Figure 1 As shown in the figure, a road surface feature recognition method for vehicles with electromechanical brakes according to an embodiment of the present invention includes the following steps: Determine whether the clamping force sensors of the electromechanical brakes of each wheel are working properly. If the clamping force sensors are working properly, the brake clamping force is directly obtained as the final brake clamping force. If the clamping force sensors fail, the first brake clamping force is obtained through the motor model, the second brake clamping force is obtained through the brake friction pad model, and the final brake clamping force is obtained by fusing them through the fuzzy logic weight algorithm. After obtaining the final brake clamping force, it is determined whether the wheel is at risk of locking up. If there is no risk of locking up, the current ground longitudinal force of the wheel is calculated based on the wheel's dynamic model. If there is a risk of locking up, the longitudinal force locking logic is triggered, and the current locked ground longitudinal force is obtained. The obtained locked ground longitudinal force is used as the current ground longitudinal force of the wheel. The current coefficient of adhesion for each wheel is calculated based on the longitudinal ground force and the final brake clamping force of each wheel. Based on the adhesion coefficient of each wheel, fuzzy rules are used to determine the type of road surface the vehicle is on.

[0024] The process of obtaining the clamping force of the first brake using the aforementioned motor model includes the following steps: A motor model based on the torque of a permanent magnet synchronous motor (PMSM) is constructed, and its voltage and torque equations are as follows: ; In the formula, For torque, , The dq-axis current of the permanent magnet synchronous motor is obtained by collecting the motor's current characteristics. , For the dq axis inductance of a permanent magnet synchronous motor. For extreme logarithms, For permanent magnet flux linkage; For surface-mounted permanent magnet synchronous motors, assuming ,but: ; Ignoring Coulomb friction and static friction torque of the motor, the equation for the electromagnetic torque of the permanent magnet synchronous motor is: ; In the formula, For the moment of inertia of the wheel, The damping coefficient is... This refers to the load torque on the motor rotor, i.e., the ball screw resistance. For mechanical angular velocity, Let angular acceleration be the angular velocity, which is obtained by differentiating the rotational process of the motor. and angular acceleration ; The ball screw, in conjunction with the reduction mechanism, converts the motor torque into axial clamping force. The clamping force of the first brake can then be derived from the motor dynamics equations. The calculation equation is: ; In the formula, The gear ratio of the reduction mechanism. For the lead screw; Then, the clamping force of the first brake is obtained by sorting. The calculation equation is: .

[0025] The process of obtaining the second brake clamping force using the aforementioned brake friction pad model includes the following steps: A brake friction pad model based on friction pad deformation is constructed. Considering that the electromechanical brake includes two stages when applying braking force: brake gap elimination and clamping force following, the total angular displacement of the motor during the braking process is first decomposed: ; In the formula, This refers to the total angular displacement of the motor during the braking process. To eliminate angular displacement due to mechanical backlash, This refers to the angular displacement of the friction plate during elastic deformation. The second brake clamping force generated by the deformation of the friction pad The nonlinear relationship that is satisfied is: ; In the formula, , , and The nonlinear stiffness coefficient is obtained through fitting via bench tests.

[0026] The process of obtaining the final brake clamping force through fuzzy logic weighting algorithm includes the following steps: The final brake clamping force is obtained by fusing the results using a fuzzy logic weighting algorithm. for: ; in, As a weighting factor, Take [0,1], when the weight factor At that time, complete trust ,when At that time, complete trust .

[0027] Among them, the aforementioned weighting factors The determination includes the following steps: In low-load scenarios, electromagnetic torque estimation is preferred, mainly because the internal friction of the mechanical structure is lower and the transmission is smoother. In high-load scenarios, displacement estimation is preferred because the internal pressure of the ball screw is high, and the deformation can well reflect the clamping force value. Under different conditions... Dynamic adjustments are required; here, a fuzzy control logic is used to adjust the weighting factors. Perform adaptive estimation; The preset maximum clamping force is Generally speaking, passenger cars The clamping force residual is defined at around 30kN. The maximum residual value is limited to ; Preset a fixed cycle and define the clamping force of the previous cycle. The membership function is: ; ; ; In the formula, These represent the clamping force in the previous cycle. The levels are low, medium, and high; 'exp()' is the exponential function. These correspond to the low, medium, and high clamping forces of the previous cycle, respectively. The proportional coefficient of the membership function can be preset based on experience; Define clamping force residual The membership function is: ; ; ; In the formula, These represent the clamping force residuals, respectively. The three levels are small, medium, and large; 'exp()' is the exponential function. These correspond to the small, medium, and large clamping force residuals, respectively. The proportional coefficient of the membership function can be preset based on experience; Define weighting factors The membership function is: ; ; ; In the formula, They represent the weighting factors respectively. The displacement-dominated, equilibrium, and torque-dominated states are represented by 'exp()', which is an exponential function. These correspond to three weighting factors: displacement-dominant, equilibrium-dominant, and torque-dominant. The proportional coefficient of the membership function can be preset based on experience; Establish weighting factors The fuzzy rule base, based on the physical property that "the greater the clamping force, the more reliable the displacement estimation," is designed with the following rules: R1: Clamping force in the previous cycle The grade is low, and the clamping force residual e is small, with a weighting factor. Tendency towards equilibrium; R2: When the clamping force in the previous cycle... When the grade is low and the clamping force residual e is medium, the weighting factor... Tendency towards torque dominance; R3: When the clamping force of the previous cycle... When the grade is low and the clamping force residual e is high, the weighting factor... Tendency towards torque dominance; R4: When the clamping force of the previous cycle... The grade is medium, and the clamping force residual e is small, with a weighting factor. Tendency towards equilibrium; R5: When the clamping force of the previous cycle... When the grade is medium and the clamping force residual e is also medium, the weighting factor... Tendency towards equilibrium; R6: When the clamping force of the previous cycle... When the grade is medium and the clamping force residual e is large, the weighting factor... Tendency towards torque dominance; R7: When the clamping force of the previous cycle... The grade is high, and the clamping force residual e is small, with a weighting factor. Prefers displacement-driven; R8: When the clamping force of the previous cycle... When the grade is high and the clamping force residual e is medium, the weighting factor... Prefers displacement-driven; R9: When the clamping force of the previous cycle... When the grade is high and the clamping force residual e is large, the weighting factor... Prefers displacement-driven; Next, the activation strength of the rule is calculated, for the first... The rules are used to calculate the activation strength. : ; in, This indicates that the smaller value on either side of the sign is taken as the result of the operation. Indicates the first The clamping force in the previous cycle of the rule. membership function, The range of values ​​is , Indicates the first Clamping force residual in the rule of the clause membership function, The range of values ​​is ; Aggregate output fuzzy set: ; in, This indicates taking a fuzzy OR of the results of the nine rules; in engineering, the maximum value (Max) is often used to achieve this. Indicates according to the first Weighting factors determined by the rules The corresponding membership function output, The range of values ​​is ; Finally, the centroid method is used for deblurring, and the weight factors are... Divide the universe of discourse [0,1] into 100 discrete points and calculate the weighted sum: ; In the formula, For the first discrete points, ; Determine the weighting factors based on the calculation results. .

[0028] The determination of whether the wheel is at risk of locking up includes the following steps: Introduce a death trend index It is used to assess the changing trend of wheel angular deceleration; when If the wheel is locked, it is determined that there is a risk of wheel lock-up; otherwise, there is no risk of wheel lock-up. This represents the rate of change of wheel angular deceleration.

[0029] If there is no risk of wheel lock-up, the longitudinal force on the ground of the wheel is calculated based on the wheel's dynamic model, including the following steps: Based on the final brake clamping force The braking torque acting on the wheel was calculated. : ; in, The coefficient of friction of the brake disc friction surface. The effective braking radius of the brake disc; Based on the longitudinal dynamics equation, the braking torque on the wheel is obtained. : ; in, For the moment of inertia of the wheel, Let be the angular deceleration of the wheel. Here, we assume the wheel deceleration is positive and write the equation accordingly. The angular deceleration of the wheel can be obtained by differentiating the wheel speed, which is measured by the onboard wheel speed sensor. The longitudinal force exerted on the wheel by the ground. The rolling radius of the wheel; Then the longitudinal force on the ground is obtained. Clamping force of the final brake The relationship is: .

[0030] If there is a risk of locking up, the longitudinal force locking logic will be triggered, and the current locked ground longitudinal force will be obtained, including the following steps: When a wheel locks up, it cannot be simply assumed that the brake clamping force and the longitudinal force of the wheel satisfy the above relationship, because the final brake clamping force of the electromechanical brake... and wheel angular deceleration It can be quickly obtained through the vehicle communication bus, so it is necessary to take appropriate measures before the wheels lock up to ensure that a reasonable longitudinal force value of the wheels can be calculated.

[0031] When a risk of wheel lockup is detected, the longitudinal force locking logic is triggered, and the initial wheel speed at the current moment is recorded. And continuously monitor wheel speed When the wheel speed decreases to a preset threshold wheel speed hour, It is assumed that the wheel has entered a deep slip phase, at which point the longitudinal force on the ground... As it approaches the road surface adhesion limit, the final brake clamping force is then determined. Calculate the longitudinal force locking the ground. for: ; In the formula, The wheel speed is The angular deceleration of the wheel at that time, The coefficient of friction of the brake disc friction surface. The effective braking radius of the brake disc. For the moment of inertia of the wheel, Let be the rolling radius of the wheel.

[0032] The vertical force obtained from the combined braking deceleration, as described above, is used to calculate the current adhesion coefficient of each wheel, including the following steps: The formula for calculating the coefficient of adhesion for wheels is: ; In the formula, It is a vertical force; Based on the linear two-degree-of-freedom model of the vehicle, the height of the center of mass is set to... The wheelbase between the front and rear axles is set to , In the formula, This is the distance from the vehicle's center of gravity to the front axle. This is the distance from the vehicle's center of gravity to the rear axle. During braking, due to inertial forces, the load on the front axle increases while the load on the rear axle decreases. Therefore, the vertical load on the front axle... and vertical load on the rear axle They are respectively: ; ; in, Let be the vehicle mass, 'a' be the vehicle braking deceleration, and 'g' be the acceleration due to gravity. Based on the longitudinal force on the ground of each wheel That is, the longitudinal force on the left front of the ground. Longitudinal force on the right front ground Longitudinal force on the left rear ground Longitudinal force on the right rear ground ,in, , , , The markings for the left front wheel, right front wheel, left rear wheel, and right rear wheel are respectively, and the coefficient of friction for the left front wheel is obtained. Right front utilizes adhesion coefficient Left and rear use adhesion coefficient Right rear utilizes adhesion coefficient for: ; When the final brake clamping force If the clamping force can be read directly from the clamping force sensor and the longitudinal force locking logic is not triggered, then the longitudinal force on the ground for each wheel will be used as the basis for the calculation. The final brake clamping force of each wheel That is, the clamping force of the left front final brake. Right front final brake clamping force Left rear final brake clamping force Right rear final brake clamping force The final expression for the adhesion coefficient of each wheel is obtained as follows: ; When the clamping force sensor fails and the longitudinal force locking logic is not triggered, the final expression for the adhesion coefficient of each wheel is obtained as follows: ; In the formula, The weighting factor for the left front wheel, The weighting factor for the right front wheel, The weighting factor for the left rear wheel. The weighting factor for the right rear wheel. The clamping force of the first brake on the left front wheel, The clamping force of the first brake on the right front wheel The clamping force of the first brake on the left rear wheel The clamping force of the first brake on the right rear wheel, For the clamping force of the second brake on the left front wheel, The clamping force of the second brake on the right front wheel The clamping force of the second brake on the left rear wheel The clamping force of the second brake on the right rear wheel; When the longitudinal force lock-up logic is triggered, the angular deceleration of the left front wheel in the final expression of the adhesion coefficient will be used to lock the aforementioned wheels. Angular deceleration of the right front wheel Angular deceleration of the left rear wheel Angular deceleration of the right rear wheel Replace with the left front wheel speed as the threshold. angular deceleration The right front wheel speed is a threshold value. angular deceleration The left rear wheel speed is a threshold value. angular deceleration The right rear wheel speed is a threshold value. angular deceleration .

[0033] The aforementioned method, which uses the adhesion coefficient of each wheel and applies fuzzy rules to determine the road surface type of the vehicle, thus providing a road surface condition basis for the vehicle stability control system, includes the following steps: Limiting the use of the adhesion coefficient on the left front Right front utilizes adhesion coefficient Left and rear use adhesion coefficient Right rear utilizes adhesion coefficient The value is in the range of [0.1, 0.8]. When the calculated wheel adhesion coefficient is less than 0.1, it is calculated as 0.1. When the calculated wheel adhesion coefficient is greater than 0.8, it is calculated as 0.8. Based on the adhesion coefficients of the four wheels, three core characteristics are calculated: ; ; ; in, The overall average of the adhesion coefficients is used for all four wheels. This represents the difference in the coefficient of adhesion between the left and right wheels. The difference in the coefficient of adhesion between the front and rear axle wheels; Next to , and Normalization is performed: ; ; ; Define the difference variables The fuzzy set is {small( ),middle( ),big( )}, and determine the difference variables. The membership function is: ; ; ; in, , All used difference variables Membership function, that is, the above Replace with ,for The membership function, which will be used to define the above... Replace with ,for Membership function; Define the mean variable The fuzzy set is {low ( ),middle( ),high( )}, and determine its membership function as: ; ; ; Determine road surface characteristics The fuzzy set is {uniform low ( ), all in one middle ( ), uniformly high ( ), double ( ), docking ( ),mix( Road surface features The value range of is [1, 7], and its membership function is determined as follows: ; ; ; ; ; ; Based on the road adhesion coefficient characteristics of each wheel, the fuzzy rules are formulated as follows: R1: If yes , yes and yes ,but It is uniformly low; R2: If yes , yes and yes ,but It is uniform; R3: If yes , yes and yes ,but It is uniformly tall; R4: If yes , yes and yes ,but It is split in half; R5: If yes , yes and yes ,but It is split in half; R6: If yes , yes and yes ,but It's about docking; R7: If yes , yes and yes ,but It's about docking; R8: If yes , yes and yes ,but It is a mixture; R9: If yes , yes and yes ,but It is a mixture; Using a fuzzy inference mechanism, inference is performed based on fuzzy input and fuzzy rules. For each fuzzy rule, a product method is applied to calculate the... Corresponding activation intensity: ; in, They are the first The corresponding rule , , membership function, The range of values ​​is , The range of values ​​is , The range of values ​​is Then, based on the activation intensity Aggregate and output fuzzy sets: ; in, For the first The output road surface features corresponding to each rule membership function, The range of values ​​is , , , , and ; The fuzzy representation of the road surface adhesion coefficient characteristics obtained from fuzzy inference is converted into specific numerical values, and the discrete centroid method is used for defuzzification to transform the road surface features. Divide the range [1,7] into 100 discrete points and calculate the weighted sum: ; In the formula, For the first Road surface features discrete points; Received Using the floor function, if the result is greater than 6, it is calculated as 6; if the result is less than 1, it is calculated as 1, and the final result is output. Road surface features The fuzzy set {uniform low ( ), all in one middle ( ), uniformly high ( ), double ( ), docking ( ),mix( The elements in} correspond to the integers 1, 2, 3, 4, 5, and 6 respectively. Based on the integer corresponding to the final result, the corresponding element is matched to confirm the current road condition.

[0034] Based on the inherent characteristics of EMB, this invention provides a road feature recognition method that maintains high accuracy even under sensor failure and wheel lock-up risks, without the need for additional sensors, thus meeting the real-time and reliability requirements of vehicle dynamic control.

[0035] The above are merely preferred embodiments of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that those skilled in the art will understand that various changes, modifications, substitutions, refinements, and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and these variations should be considered within the scope of protection of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for road surface feature recognition of vehicles based on electromechanical brakes, characterized in that, Includes the following steps: Determine whether the clamping force sensors of the electromechanical brakes of each wheel are working properly. If the clamping force sensors are working properly, the brake clamping force is directly obtained as the final brake clamping force. If the clamping force sensors fail, the first brake clamping force is obtained through the motor model, the second brake clamping force is obtained through the brake friction pad model, and the final brake clamping force is obtained by fusing them through the fuzzy logic weight algorithm. After obtaining the final brake clamping force, it is determined whether the wheel is at risk of locking up. If there is no risk of locking up, the current ground longitudinal force of the wheel is calculated based on the wheel's dynamic model. If there is a risk of locking up, the longitudinal force locking logic is triggered, and the current locked ground longitudinal force is obtained. The obtained locked ground longitudinal force is used as the current ground longitudinal force of the wheel. The current coefficient of adhesion for each wheel is calculated based on the longitudinal ground force and the final brake clamping force of each wheel. Based on the adhesion coefficient of each wheel, fuzzy rules are used to determine the type of road surface the vehicle is on.

2. The method for road surface feature recognition of a vehicle based on electromechanical braking according to claim 1, characterized in that, The process of obtaining the clamping force of the first brake through the motor model includes the following steps: A motor model based on the torque of a permanent magnet synchronous motor is constructed, and the clamping force of the first brake is calculated. : ； In the formula, This represents the q-axis current of the permanent magnet synchronous motor. For extreme logarithms, It is a permanent magnet flux chain. For the moment of inertia of the wheel, The damping coefficient is... For mechanical angular velocity, Angular acceleration, The gear ratio of the reduction mechanism. This refers to the lead of the leadscrew.

3. The method for road feature recognition of vehicles based on electromechanical brakes according to claim 2, characterized in that, The process of obtaining the second brake clamping force through the brake friction pad model includes the following steps: A brake friction pad model based on friction pad deformation is constructed, and the total angular displacement of the motor during braking is decomposed to obtain the angular displacement that eliminates mechanical backlash. Angular displacement of friction plate undergoing elastic deformation And calculate the second brake clamping force generated by the deformation of the friction pad. for: ； In the formula, , , and This is the nonlinear stiffness coefficient.

4. The method for road surface feature recognition of a vehicle based on electromechanical braking according to claim 3, characterized in that, The process of obtaining the final brake clamping force through fuzzy logic weighting algorithm includes the following steps: The final brake clamping force is obtained by fusing the results using a fuzzy logic weighting algorithm. for: In the formula, As a weighting factor, Take [0,1].

5. A method for road surface feature recognition of a vehicle based on electromechanical braking according to claim 4, characterized in that, The weighting factor The determination includes the following steps: The preset maximum clamping force is Define clamping force residual The maximum residual value is limited to ; Preset a fixed cycle and define the clamping force of the previous cycle. The three membership functions , and , and These represent the clamping force in the previous cycle. The clamping force residual is defined as low, medium, and high. The three membership functions , and , and These represent the clamping force residuals, respectively. Three levels: small, medium, and large; defining weighting factors. The three membership functions , and , , and They represent the weighting factors respectively. The three degrees of displacement dominance, equilibrium, and torque dominance; Established Weighting factors of the rules The fuzzy rule base, then for the first The rules are used to calculate the activation strength. : ； In the formula, This indicates that the smaller value on either side of the sign is taken as the result of the operation. Indicates the first The clamping force in the previous cycle of the rule. membership function, The range of values ​​is and , Indicates the first Clamping force residual in the rule of the clause membership function, The range of values ​​is and ; Aggregate the output fuzzy set and use the centroid method for defuzzification, then adjust the weight factors. The domain is divided into 100 discrete points on the domain [0,1], and a weighted sum is calculated. The weighting factors are determined based on the calculation results. : ； ； In the formula, Indicates to The result of this rule is taken as "fuzzy OR". Indicates according to the first Weighting factors determined by the rules The corresponding membership function output, The range of values ​​is and In the formula, For the first A discrete point.

6. The method for road surface feature recognition of a vehicle based on electromechanical braking according to claim 1, characterized in that, The process of determining whether a wheel is at risk of locking up includes the following steps: Introduce a death trend index It is used to assess the changing trend of wheel angular deceleration; when If the wheel is locked, it is determined that there is a risk of wheel lock-up; otherwise, there is no risk of wheel lock-up. This represents the rate of change of wheel angular deceleration.

7. The method for road surface feature recognition of a vehicle based on electromechanical braking according to claim 1, characterized in that, If there is no risk of wheel lock-up, the longitudinal force on the ground of the wheel is calculated based on the wheel's dynamic model, including the following steps: Based on the dynamic model of the wheel, the longitudinal force on the ground is obtained. Clamping force of the final brake The relationship is: ； In the formula, The coefficient of friction of the brake disc friction surface. The effective braking radius of the brake disc. For the moment of inertia of the wheel, The angular deceleration of the wheel, The longitudinal force exerted on the wheel by the ground. Let be the rolling radius of the wheel.

8. The method for road surface feature recognition of a vehicle based on electromechanical braking according to claim 1, characterized in that, If there is a risk of locking up, the longitudinal force locking logic is triggered, and the current locked ground longitudinal force is obtained, including the following steps: When a risk of wheel lockup is detected, the longitudinal force locking logic is triggered, and the initial wheel speed at the current moment is recorded. And continuously monitor wheel speed When the wheel speed decreases to a preset threshold wheel speed At this point, it is considered that the wheel has entered the deep slip phase, at which time the longitudinal force on the ground is... As it approaches the road surface adhesion limit, the final brake clamping force is then determined. Calculate the longitudinal force locking the ground. for: ； In the formula, The wheel speed is The angular deceleration of the wheel at that time, The coefficient of friction of the brake disc friction surface. The effective braking radius of the brake disc. For the moment of inertia of the wheel, Let be the rolling radius of the wheel.

9. A method for road surface feature recognition of a vehicle based on electromechanical brakes according to claim 1, characterized in that, The calculation of the current coefficient of adhesion for each wheel based on the longitudinal force on the ground and the final brake clamping force includes the following steps: Based on the linear two-degree-of-freedom model of the vehicle, the height of the center of mass is set to... The wheelbase between the front and rear axles is set to , In the formula, This is the distance from the vehicle's center of gravity to the front axle. This is the distance from the vehicle's center of gravity to the rear axle. Based on the longitudinal force on the ground of each wheel That is, the longitudinal force on the left front of the ground. Longitudinal force on the right front ground Longitudinal force on the left rear ground and right rear longitudinal force And the final brake clamping force of each wheel. That is, the clamping force of the left front final brake. Right front final brake clamping force Left rear final brake clamping force and right rear final brake clamping force The adhesion coefficients of the left front were obtained respectively. Right front utilizes adhesion coefficient Left and rear use adhesion coefficient And right rear use of adhesion coefficient : ； In the formula, The coefficient of friction of the brake disc friction surface. Let be the vehicle mass, 'a' be the vehicle braking deceleration, and 'g' be the acceleration due to gravity. The effective braking radius of the brake disc. The rolling radius of the wheel, For the moment of inertia of the wheel, The angular deceleration of the left front wheel, The angular deceleration of the right front wheel. The angular deceleration of the left rear wheel, This is the angular deceleration of the right rear wheel.

10. A method for road surface feature recognition of a vehicle based on electromechanical braking according to claim 9, characterized in that, The method of determining the road surface type based on the adhesion coefficient of each wheel using fuzzy rules to provide road condition information for the vehicle stability control system includes the following steps: Based on the adhesion coefficient of the left front wheel... Right front utilizes adhesion coefficient Left and rear use adhesion coefficient And right rear use of adhesion coefficient Calculate three core features: ； ； ； In the formula, The overall average of the adhesion coefficients is used for all four wheels. This represents the difference in the coefficient of adhesion between the left and right wheels. The difference in the coefficient of adhesion between the front and rear axle wheels; Next to , and Normalization is performed to obtain the corresponding result. , and ; Define the difference variables The fuzzy set is {small( ),middle( ),big( )}, and determine three levels of difference variables. The membership function, and and All used difference variables Membership function; defining the mean variable The fuzzy set is {low( ),middle( ),high( )}, and determine the mean variable for three degrees. Membership function; define road surface features The fuzzy set is {uniform low ( ), all in one middle ( ), uniformly high ( ), double ( ), docking ( ),mix( )}, and identified six levels of road surface characteristics. The membership function, and at the same time, the road surface features The elements in the fuzzy set correspond to the integers 1, 2, 3, 4, 5, and 6, respectively. Based on the road adhesion coefficient characteristics of each wheel, formulate The first fuzzy rule is determined; using a fuzzy inference mechanism, inference is performed based on the fuzzified input and fuzzy rules, and for each fuzzy rule, a product method is used to calculate the first fuzzy rule. The activation intensity corresponding to each fuzzy rule: ； In the formula, and They are the first The corresponding fuzzy rules , and membership function, The range of values ​​is and , The range of values ​​is and , The range of values ​​is and ; Based on activation intensity Aggregate the output fuzzy set to include road surface features Divide the range [1,7] into 100 discrete points and calculate the weighted sum; the resulting... The system uses a floor function to round down and outputs the final result. The final result is then mapped to road surface features. The integers corresponding to the elements in the fuzzy set are used to determine the current road surface conditions: ； ； In the formula, For the first The output road surface features corresponding to each rule membership function, The range of values ​​is , , , , and ; For the first Road surface features The discrete points.

Images