Hydraulic pressure control method for a brake-by-wire system

Abstract

The application relates to the field of automobile brake-by-wire, and specifically discloses a brake-by-wire system hydraulic pressure control method, which comprises the following steps: establishing a hydraulic dynamics model of the brake-by-wire system, adopting a Lugre dynamic friction model to describe pipeline friction behavior, introducing brake fluid temperature, vehicle longitudinal acceleration and tire adhesion coefficient as dynamic input parameters, and constructing a pressure transmission equation; based on the hydraulic dynamics model, a compound control architecture of feedforward compensation plus feedback regulation is designed; an improved particle swarm algorithm is used to optimize the extended state observer gain and nonlinear error feedback control law parameters of an active disturbance rejection controller; wheel cylinder pressure, brake motor current, displacement sensor signals and vehicle state parameters of the brake-by-wire system are collected in real time, the system total disturbance is estimated by the extended state observer of the active disturbance rejection controller and compensated in real time, and a wheel cylinder hydraulic pressure control instruction is output. The application improves the hydraulic pressure control precision, robustness and working condition adaptability, and guarantees brake stability.

Description

Technical Field

[0001] This invention relates to the field of automotive brake-by-wire technology, specifically to a hydraulic pressure control method for a brake-by-wire system. Background Technology

[0002] As a core chassis component of intelligent connected vehicles and new energy vehicles, the brake-by-wire system directly determines the vehicle's braking safety, stability, and comfort. Its hydraulic pressure control precision and robustness are crucial for ensuring driving safety. With the development of autonomous driving technology, vehicles place higher demands on the dynamic response speed, nonlinear operating condition adaptability, and fault tolerance of the braking system. Precise hydraulic pressure control can effectively avoid risks such as wheel lock-up and brake deviation, while simultaneously balancing braking performance and driving experience, playing an irreplaceable role in complex scenarios such as emergency braking and cornering braking.

[0003] Currently, several mature technologies have been developed for hydraulic pressure control in automotive braking systems, achieving pressure regulation through the construction of hydraulic dynamics models combined with control algorithms. Among these, PID control methods and their improved versions are widely used in traditional braking systems due to their simple structure and ease of engineering implementation. Model predictive control (MRC) schemes improve overall performance through multi-objective optimization. Adaptive control and sliding mode control are also commonly used to address system parameter drift and external disturbances. To enhance reliability, some solutions employ hardware redundancy designs with dual sensors and dual actuators, achieving initial fault tolerance through signal verification. Furthermore, addressing the nonlinear characteristics of braking systems, some technologies attempt to correct errors through linearization approximations or simple compensation strategies, combined with algorithms such as recursive least squares to update model parameters.

[0004] However, existing technologies still have many shortcomings in practical applications: First, the modeling accuracy of nonlinear characteristics is insufficient. Nonlinear factors such as brake line friction, hydraulic valve dead zone, and brake fluid compressibility are mostly linearized, making it difficult to accurately capture the frictional abrupt changes under low-speed conditions and the dynamic coupling relationship under high-speed conditions, leading to pressure overshoot or response delay. Second, the adaptability of control algorithms to different operating conditions is limited. General control parameters cannot match the performance requirements of different scenarios such as emergency braking, hill braking, and cornering braking. Furthermore, traditional optimization algorithms are prone to getting trapped in local optima, making it difficult to balance pressure tracking accuracy, response speed, and overshoot. Third, robustness and real-time performance are also issues. There are contradictions: while complex model predictive control can optimize multi-objective performance, it requires a large amount of computation and cannot meet the millisecond-level response requirements, while simple control algorithms have weak ability to suppress disturbances such as brake fluid temperature changes and pipeline leaks; fourth, the redundancy design is imperfect, mostly concentrated at the hardware level, lacking software parsing redundancy and fault classification fault tolerance strategies, and the fault diagnosis and switching logic is simple, making it difficult to deal with different types of faults such as sensor drift and actuator jamming; fifth, the coordination with vehicle dynamics is insufficient, and the state information such as steer-by-wire and vehicle attitude is not fully integrated, resulting in low matching degree between hydraulic pressure distribution and vehicle dynamics, affecting braking stability under complex working conditions.

[0005] The aforementioned deficiencies are particularly pronounced under complex road conditions and dynamic loads, severely restricting the performance improvement of brake-by-wire systems and failing to meet the high safety, high precision, and high reliability requirements of intelligent vehicles for braking systems. Therefore, it is urgent to address the shortcomings of existing technologies in nonlinear compensation, operating condition adaptation, and disturbance suppression by optimizing hydraulic dynamics modeling methods, innovating control architectures and algorithms, and improving redundancy and fault-tolerant design. This will enable precise, rapid, and stable control of hydraulic force, providing a more reliable technical guarantee for vehicle braking safety. Summary of the Invention

[0006] To address the aforementioned problems in the prior art, this invention provides a hydraulic pressure control method for a brake-by-wire system, which solves the problems of insufficient nonlinear compensation, poor adaptability to operating conditions, and imbalance between robustness and real-time performance in the prior art, thereby improving braking control accuracy and safety.

[0007] To achieve the above objectives, this invention proposes a hydraulic pressure control method for a brake-by-wire system, comprising the following steps: S1. Establish a hydraulic dynamics model for the brake-by-wire system. This model integrates the nonlinear characteristics of brake line friction and hydraulic valve dead zone, and uses the Lugre dynamic friction model to describe the line friction behavior. It also incorporates brake fluid temperature T and vehicle longitudinal acceleration. Tire adhesion coefficient As dynamic input parameters, pressure transmission equations are constructed; S2. Based on the hydraulic dynamics model, a composite control architecture of feedforward compensation and feedback regulation is designed. The feedforward compensation module generates compensation torque according to the output results of the Lugre dynamic friction model, and the feedback regulation module uses an active disturbance rejection controller (ADRC) to track the target hydraulic pressure. S3. The extended state observer (ESO) gain and nonlinear error feedback control law parameters of the active disturbance rejection controller are optimized by improving the particle swarm optimization (PSO) algorithm. The improved PSO algorithm adopts a dynamic learning factor and nonlinear inertial weight adjustment strategy to obtain the objective function of the improved PSO algorithm. S4. Real-time acquisition of wheel cylinder pressure, brake motor current, displacement sensor signals and vehicle status parameters of the brake-by-wire system; estimation of total system disturbance and real-time compensation through the extended state observer of the active disturbance rejection controller; output of wheel cylinder hydraulic pressure control command.

[0008] Preferably, in S1, the hydraulic dynamics model integrates the nonlinear characteristics of brake line friction and hydraulic valve dead zone, and uses the Lugre dynamic friction model to describe the line friction behavior. Its complete expression is: ; ; ; when hour, ; In the formula, For the frictional torque of the braking pipeline, The average stiffness of the bristles. The viscous damping coefficient is... This is the bristle damping coefficient. This represents the average deformation of the bristles on the contact surface. The relative velocity of the pipeline, The rate of change of the average deformation of the bristles. For Stribeck characteristic function, For Coulomb friction torque, For the maximum static friction torque, Stribeck characteristic velocity, It is a natural constant. This is a minimum speed threshold used to avoid numerical singularities; The constructed pressure transmission equation is: ; In the formula, For the wheel cylinder pressure, For hydraulic valve flow gain, For effective control of the voltage of the hydraulic valve, The total leakage coefficient of the system. For the effective area of ​​the wheel cylinder, The speed of the piston movement in the cylinder. This refers to the total volume of the hydraulic system. This is the effective volumetric elastic modulus of the brake fluid.

[0009] Preferably, in S1, the process of establishing the hydraulic dynamics model of the brake-by-wire system further includes establishing a nonlinear model of the valve orifice dead zone by combining the flow-pressure characteristic curve of the brake valve, and characterizing the pressure response characteristics within the dead zone range by a piecewise function, wherein the piecewise function is characterized as follows: ; In the formula, For effective control of the voltage of the hydraulic valve, The original control voltage, This is the dead zone threshold.

[0010] Preferably, in S3, the expression for the dynamic learning factor is: ; ; The expression for the nonlinear inertia weight is as follows: ; In the formula, This represents the maximum value of the individual learning factor. This represents the minimum value of the individual learning factor. This represents the maximum value of the social learning factor. This represents the minimum value of the social learning factor. This represents the current iteration number. The maximum number of iterations, This is the adjustment coefficient for individual learning factors. The social learning factor adjustment coefficient. This represents the maximum value of the inertia weight. This represents the minimum value of the inertia weight. The objective function of the improved particle swarm optimization algorithm is: ; In the formula, For the target hydraulic pressure, This is the actual hydraulic pressure. For pressure response time, For overshoot, The weighting coefficients for pressure tracking accuracy are... The pressure response speed weighting coefficient is used. This is a weighting factor for pressure overshoot, and the sum of these factors must be 1. The weighting factor is dynamically adjusted according to the braking conditions, especially during emergency braking. During normal braking , For integration time; The extended state observer (ESO) gain of the active disturbance rejection controller adopts a nonlinear gain function, and the ESO gain expression is: ; ; ; In the formula, This is a pressure estimate. This is an estimate of the rate of change of pressure. This is the estimated total disturbance. For measurement output, To control the gain, , , , For the observer bandwidth, This is a nonlinear factor used to adjust the feedback nonlinearity of the pressure estimate, with a value ranging from 0.8 to 1. This is a nonlinear factor used to adjust the feedback nonlinearity of the pressure change rate estimate, with a value ranging from 0.5 to 0.8. This is a nonlinear factor used to adjust the feedback nonlinearity of the total disturbance estimate, with a value ranging from 0.1 to 0.5, and satisfying the following conditions: , For nonlinear error functions: ; In the formula, For observation error, This is the threshold for the linear interval.

[0011] Preferably, the method further includes a working condition identification and parameter self-tuning step: Braking conditions are identified using the K-Nearest Neighbors (KNN) algorithm, including emergency braking, normal deceleration, hill braking, and cornering braking; a preset parameter set is called for different braking conditions, and the key parameters of the hydraulic dynamics model are updated in real time using the Recursive Least Squares (RLS) method, including the pipeline friction coefficient and the brake fluid viscosity coefficient. The RLS algorithm iterative formula is: ; ; ; in, For the first Time parameter estimates, For recursive gain, For measurement output, For the regression vector, Let be the error covariance matrix. The forgetting factor has a value range of (0.95, 0.99).

[0012] Preferably, it also includes a hardware / software integrated solution for redundancy design: A dual-path pressure acquisition mechanism with a main channel and a backup channel is constructed. The main channel directly collects the wheel cylinder pressure through a pressure sensor. The backup channel is based on the brake motor current. With displacement sensor signal Establish a pressure estimation model: ; in, For cylinder wheel pressure estimation, This refers to the motor current-pressure gain coefficient. This is the piston speed-pressure dynamic coefficient. This is the static coefficient of piston displacement-pressure. The model bias is obtained through offline calibration. The residual chi-square test algorithm is used to detect the sensor status in real time. When the main channel sensor fails, it automatically switches to the backup channel to obtain the pressure feedback signal. The switching process is imperceptible.

[0013] Preferably, the implementation process of the residual chi-square test algorithm includes: calculating the residual sequence between the measured pressure value of the main channel and the estimated pressure value of the backup channel. ; The residual sequence follows a mean of 0 and a variance of . The normal distribution; Construct the residual function and make a judgment; when the residual function value is greater than the degrees of freedom... Confidence level is The critical value of the chi-square distribution, i.e. When a main channel sensor malfunction is detected, channel switching is triggered. The residual function is: ; In the formula, This is the length of the sliding window.

[0014] Preferably, in S4, the vehicle state parameters also include the wheel angles of the steer-by-wire system. yaw rate of the vehicle body Based on the vehicle's overall state parameters, the wheel cylinder hydraulic pressure distribution is performed. The formula for the left and right wheel pressure distribution during cornering braking is as follows: ; ; ; ; In the formula, For the left wheel cylinder pressure, For the right wheel cylinder pressure, The distribution coefficient takes the value (0, 1). This refers to the load transfer amount between the left and right wheels. For the vertical load on the tire, For the overall vehicle quality, It is lateral acceleration. For the height of the center of mass, The wheelbase is the distance between the wheels. For vehicle speed, This refers to the vehicle's longitudinal acceleration.

[0015] Preferably, it also includes a fault classification and fault tolerance step: A preset fault level is defined, including sensor drift fault, actuator partial failure fault, and minor pipeline leakage fault. In the case of sensor drift fault, a moving average filtering algorithm is used to correct the measured value. ; In the formula, This is the length of the filtering window; In the event of a partial actuator failure, the system switches to the backup execution channel; in the event of a minor pipeline leak, the control gain is increased through the disturbance compensation module of the active disturbance rejection controller, and the corrected control law is as follows: ; In the formula, The original control signal, To compensate for leakage gain, This is an estimate of the leakage disturbance.

[0016] Preferably, it also includes an adaptive compensation step for the adhesion coefficient: Based on wheel speed and wheel cylinder pressure signals, and combined with the Magic Tire model, the tire-road adhesion coefficient is estimated. The expression for the lateral force of the Magic Tire model is as follows: In the formula, This refers to the lateral force of the tire. Side slip angle, Stiffness factor For shape factor, As the peak factor, Curvature factor; The precise coefficient of adhesion of the tire is obtained by recursively solving using the magic tire model. , The upper limit of the wheel cylinder hydraulic pressure is dynamically adjusted. ,when At that time, reduce the upper limit of hydraulic pressure to , This is a preset threshold.

[0017] Therefore, this invention proposes a hydraulic pressure control method for a brake-by-wire system, which has the following advantages: (1) By integrating the Lugre dynamic friction model and hydraulic transmission equation, nonlinear factors such as pipeline friction and valve dead zone are fully compensated. Combined with multiple dynamic input parameters, pressure overshoot and response delay are significantly reduced, and accurate modeling improves control accuracy. (2) By improving PSO to optimize ADRC parameters, dynamically adapt to different braking scenarios, ESO to compensate for disturbances in real time, balance control accuracy and real-time performance, and enhance working condition adaptability and robustness. (3) The software and hardware integration redundancy design and fault classification strategy, combined with the vehicle's collaborative pressure distribution, significantly improves braking stability and fault response capabilities, and redundancy and fault tolerance ensure safety and reliability.

[0018] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the overall process of a hydraulic pressure control method for a brake-by-wire system according to the present invention; Figure 2 This is a schematic diagram of the hydraulic dynamics model structure of the hydraulic pressure control method for a brake-by-wire system according to the present invention; Figure 3 This is a schematic diagram of a composite control architecture of feedforward compensation + ADRC feedback for a hydraulic pressure control method of a brake-by-wire system according to the present invention. Figure 4 This is a schematic diagram illustrating the hardware and software fusion redundancy design of the hydraulic pressure control method for a brake-by-wire system according to the present invention. Detailed Implementation

[0020] To make the technical solutions, advantages, and objectives of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below. The described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the described embodiments of the present invention without creative effort are within the protection scope of this application.

[0021] Unless otherwise defined, the technical or scientific terms used in this invention shall have the ordinary meaning as understood by one of ordinary skill in the art to which this invention pertains.

[0022] like Figures 1-4As shown, the present invention provides a hydraulic pressure control method for a brake-by-wire system, comprising the following steps: S1. Establish a hydraulic dynamics model for the brake-by-wire system. This model integrates the nonlinear characteristics of brake line friction and hydraulic valve dead zone, and uses the Lugre dynamic friction model to describe the line friction behavior. It also incorporates brake fluid temperature T and vehicle longitudinal acceleration. Tire adhesion coefficient As dynamic input parameters, pressure transmission equations are constructed; The hydraulic dynamics model integrates the nonlinear characteristics of brake line friction and hydraulic valve dead zone, and uses the Lugre dynamic friction model to describe the line friction behavior. Its complete expression is: ; ; ; when hour, ; In the formula, For the frictional torque of the braking pipeline, The average stiffness of the bristles. The viscous damping coefficient is... This is the bristle damping coefficient. This represents the average deformation of the bristles on the contact surface. The relative velocity of the pipeline, The rate of change of the average deformation of the bristles. For Stribeck characteristic function, For Coulomb friction torque, For the maximum static friction torque, Stribeck characteristic velocity, It is a natural constant. This is a minimum speed threshold used to avoid numerical singularities; The constructed pressure transmission equation is as follows: ; In the formula, For the wheel cylinder pressure, For hydraulic valve flow gain, For effective control of the voltage of the hydraulic valve, The total leakage coefficient of the system. For the effective area of ​​the wheel cylinder, The speed of the piston movement in the cylinder. This refers to the total volume of the hydraulic system. This is the effective volumetric elastic modulus of the brake fluid.

[0023] The process of establishing the hydraulic dynamics model of the brake-by-wire system also includes establishing a nonlinear model of the valve orifice dead zone by combining the flow-pressure characteristic curve of the brake valve, and characterizing the pressure response characteristics within the dead zone range through piecewise functions. The piecewise functions are characterized as follows: ; In the formula, For effective control of the voltage of the hydraulic valve, The original control voltage, This is the dead zone threshold.

[0024] S2. Based on the hydraulic dynamics model, a composite control architecture of feedforward compensation and feedback regulation is designed. The feedforward compensation module generates compensation torque according to the output results of the Lugre dynamic friction model, and the feedback regulation module uses an active disturbance rejection controller (ADRC) to track the target hydraulic pressure. S3. By improving the particle swarm optimization (PSO) algorithm, the extended state observer (ESO) gain and nonlinear error feedback control law parameters of the active disturbance rejection controller are optimized. The improved PSO algorithm adopts a dynamic learning factor and nonlinear inertial weight adjustment strategy to obtain the objective function of the improved PSO algorithm. The expression for the dynamic learning factor is: ; ; The expression for nonlinear inertia weight is: ; In the formula, This represents the maximum value of the individual learning factor. This represents the minimum value of the individual learning factor. This represents the maximum value of the social learning factor. This represents the minimum value of the social learning factor. This represents the current iteration number. The maximum number of iterations, This is the adjustment coefficient for individual learning factors. The social learning factor adjustment coefficient. This represents the maximum value of the inertia weight. This represents the minimum value of the inertia weight. The optimization objective function of the improved particle swarm optimization algorithm is: ; In the formula, For the target hydraulic pressure, This is the actual hydraulic pressure. For pressure response time, For overshoot, The weighting coefficients for pressure tracking accuracy are... The pressure response speed weighting coefficient is used. This is a weighting factor for pressure overshoot, and the sum of these factors must be 1. The weighting factor is dynamically adjusted according to the braking conditions, especially during emergency braking. During normal braking , For integration time; The extended state observer (ESO) gain of the active disturbance rejection controller adopts a nonlinear gain function, and the ESO gain expression is: ; ; ; In the formula, This is a pressure estimate. This is an estimate of the rate of change of pressure. This is the estimated total disturbance. For measurement output, To control the gain, , , , For the observer bandwidth, This is a nonlinear factor used to adjust the feedback nonlinearity of the pressure estimate, with a value ranging from 0.8 to 1. This is a nonlinear factor used to adjust the feedback nonlinearity of the pressure change rate estimate, with a value ranging from 0.5 to 0.8. This is a nonlinear factor used to adjust the feedback nonlinearity of the total disturbance estimate, with a value ranging from 0.1 to 0.5, and satisfying the following conditions: , For nonlinear error functions: ; In the formula, For observation error, This is the threshold for the linear interval.

[0025] S4. Real-time acquisition of wheel cylinder pressure, brake motor current, displacement sensor signals and vehicle status parameters of the brake-by-wire system. Estimates the total system disturbance through the extended state observer of the active disturbance rejection controller and performs real-time compensation, outputting wheel cylinder hydraulic pressure control commands.

[0026] The vehicle status parameters also include the wheel angles of the steer-by-wire system. yaw rate of the vehicle body Based on the vehicle's overall condition parameters, the wheel cylinder hydraulic pressure distribution is calculated. The formula for the left and right wheel pressure distribution during cornering braking is as follows: ; ; ; ; In the formula, For the left wheel cylinder pressure, For the right wheel cylinder pressure, The distribution coefficient takes the value (0, 1). This refers to the load transfer amount between the left and right wheels. For the vertical load on the tire, For the overall vehicle quality, It is lateral acceleration. For the height of the center of mass, The wheelbase is the distance between the wheels. For vehicle speed, This refers to the vehicle's longitudinal acceleration.

[0027] It also includes operating condition identification and parameter self-tuning steps: Braking conditions are identified using the K-Nearest Neighbors (KNN) algorithm, including emergency braking, normal deceleration, hill braking, and cornering braking; a preset parameter set is called for different braking conditions, and the Recursive Least Squares (RLS) method is used to update the key parameters of the hydraulic dynamics model in real time. These key parameters include the pipeline friction coefficient and the brake fluid viscosity coefficient. The RLS algorithm iterative formula is as follows: ; ; ; in, For the first Time parameter estimates, For recursive gain, For measurement output, For the regression vector, Let be the error covariance matrix. The forgetting factor has a value range of (0.95, 0.99).

[0028] It also includes the analytical redundancy design steps for hardware and software integration: A dual-path pressure acquisition mechanism with a main channel and a backup channel is constructed. The main channel directly collects the wheel cylinder pressure through a pressure sensor. The backup channel is based on the brake motor current. With displacement sensor signal Establish a pressure estimation model: ; in, For cylinder wheel pressure estimation, This refers to the motor current-pressure gain coefficient. This is the piston speed-pressure dynamic coefficient. This is the static coefficient of piston displacement-pressure. The model bias is obtained through offline calibration. The residual chi-square test algorithm is used to detect the sensor status in real time. When the main channel sensor fails, it automatically switches to the backup channel to obtain the pressure feedback signal. The switching process is imperceptible.

[0029] The implementation process of the residual chi-square test algorithm includes: calculating the residual sequence between the measured pressure values ​​of the main channel and the estimated pressure values ​​of the backup channel. ; The residual sequence follows a pattern with a mean of 0 and a variance of 0. The normal distribution; Construct the residual function and make a judgment; when the residual function value is greater than the degrees of freedom... Confidence level is The critical value of the chi-square distribution, i.e. When a main channel sensor malfunction is detected, channel switching is triggered, and the residual function is: ; In the formula, This is the length of the sliding window.

[0030] Preferably, it also includes a fault classification and fault tolerance step: Preset fault levels include sensor drift fault, actuator partial failure fault, and minor pipeline leakage fault; in the case of sensor drift fault, a moving average filtering algorithm is used to correct the measured value. ; In the formula, This is the length of the filtering window; In the event of a partial actuator failure, the system switches to the backup execution channel; in the event of a minor pipeline leak, the control gain is increased through the disturbance compensation module of the active disturbance rejection controller, and the corrected control law is as follows: ; In the formula, The original control signal, To compensate for leakage gain, This is an estimate of the leakage disturbance.

[0031] It also includes an adaptive compensation step for the adhesion coefficient: Based on wheel speed and wheel cylinder pressure signals, and combined with the Magic Tire model, the tire-road adhesion coefficient is estimated. The expression for the lateral force of the Magic Tire model is as follows: In the formula, This refers to the lateral force of the tire. Side slip angle, Stiffness factor For shape factor, As the peak factor, Curvature factor; The precise coefficient of adhesion of the tire is obtained by recursively solving using the magic tire model. , The upper limit of the wheel cylinder hydraulic pressure is dynamically adjusted. ,when At that time, reduce the upper limit of hydraulic pressure to , This is a preset threshold.

[0032] This invention uses the brake-by-wire system of a certain new energy pure electric passenger vehicle as the test object to verify the control accuracy, response speed and robustness of the hydraulic pressure control method under different braking conditions.

[0033] The test vehicle's brake-by-wire system is equipped with a brushless brake motor, high-precision wheel cylinder pressure sensors, displacement sensors, and a vehicle dynamics controller. The brake fluid used is DOT4 type. The dead zone threshold of the system's hydraulic valves is calibrated to 0.2V. The key parameters of the Lugre dynamic friction model were calibrated by bench testing: average bristle stiffness is 1200 N / m, viscous damping coefficient is 0.8 N·s / m, bristle damping coefficient is 0.5 N·s / m, Stribeck characteristic velocity is 0.05 m / s, maximum static friction torque is 3.2 N·m, and Coulomb friction torque is 2.8 N·m.

[0034] The test was conducted on an automotive chassis dynamometer, simulating four typical operating conditions: emergency braking, normal deceleration, hill braking (15° incline), and cornering braking (15m turning radius). Simultaneously, brake fluid temperature changes (-20℃ to 80℃), minor pipeline leaks (leakage coefficient 0.002mL / (s・kPa)), and sensor drift (drift amount) were introduced. The specific implementation process is as follows, including disturbance conditions such as 2% (2%): The test vehicle's brake-by-wire system was powered on and initialized. Sensor zero-position signals were collected and zero-point calibration was completed. The basic parameters of the hydraulic dynamics model, such as hydraulic valve flow gain, wheel cylinder effective area, and total hydraulic system volume, were calibrated through offline bench tests. At the same time, the motor current-pressure gain coefficient, piston speed-pressure dynamic coefficient, piston displacement-pressure static coefficient, and model bias of the pressure estimation model were calibrated. The iterative parameters of the improved particle swarm optimization algorithm are preset as follows: the maximum value of the individual learning factor is 2.5 and the minimum value is 0.5; the maximum value of the social learning factor is 2.0 and the minimum value is 0.4; the maximum value of the inertia weight is 0.9 and the minimum value is 0.4; the maximum number of iterations is 200; the bandwidth of the active disturbance rejection controller observer is set to 50 rad / s; the nonlinear factors are set to 0.9 (pressure estimate), 0.7 (pressure change rate estimate), and 0.3 (total disturbance estimate); and the linear interval threshold is set to 0.01.

[0035] Based on the test vehicle parameters, a hydraulic dynamics model integrating the Lugre dynamic friction model and the hydraulic valve dead zone nonlinear model was constructed. Brake fluid temperature, vehicle longitudinal acceleration, and tire adhesion coefficient were introduced as dynamic input parameters to complete the pressure transmission equation. A braking condition identification sample library was built using the K-nearest neighbor (KNN) algorithm, and the feature values ​​of the vehicle state parameters for four typical operating conditions were entered. At the same time, the recursive least squares method was started, and the forgetting factor was set to 0.97 to update key parameters such as pipeline friction coefficient and brake fluid viscosity coefficient in real time, so as to achieve self-tuning of model parameters.

[0036] With pressure tracking accuracy, response speed, and overshoot as optimization objectives, an improved particle swarm optimization algorithm is constructed, and weighting coefficients are preset according to the working conditions: during emergency braking, the weighting coefficient for pressure tracking accuracy is 0.4, the weighting coefficient for response speed is 0.4, and the weighting coefficient for overshoot is 0.2; during normal braking, hill braking, and cornering braking, the weighting coefficient for pressure tracking accuracy is 0.5, the weighting coefficient for response speed is 0.3, and the weighting coefficient for overshoot is 0.2. By using a dynamic learning factor and a nonlinear inertial weight adjustment strategy, the extended state observer gain and nonlinear error feedback control law parameters of the active disturbance rejection controller are iteratively optimized to obtain the optimal parameter set under each typical operating condition and store it in the controller.

[0037] Four typical braking conditions were simulated sequentially on a chassis dynamometer, with the test parameters set as follows: Standard deceleration: Initial vehicle speed 60 km / h, target deceleration 2 m / s 2 Brake fluid temperature 25℃; Emergency braking: Initial vehicle speed 80 km / h, target deceleration 8 m / s² 2 Brake fluid temperature 25℃; Hill Start Braking: 15° uphill slope, initial vehicle speed 30 km / h, target deceleration 3 m / s² 2 Brake fluid temperature 0℃; Curving braking: Turning radius 15m, initial vehicle speed 40km / h, target deceleration 4m / s² 2 Brake fluid temperature 40℃.

[0038] During the test, the wheel cylinder pressure, brake motor current, displacement sensor signals, and vehicle state parameters (wheel angle, body yaw rate, longitudinal or lateral acceleration, vehicle mass, center of gravity height, etc.) of the brake-by-wire system were collected in real time. The feedforward compensation module generated compensation torque based on the output results of the Lugre dynamic friction model. The extended state observer of the active disturbance rejection controller estimated the total disturbance of the system in real time and completed the compensation. The feedback adjustment module tracked the target hydraulic pressure and output control commands. Under the cornering braking condition, the load transfer of the left and right wheels was calculated based on the vehicle state parameters, and the differential distribution of hydraulic pressure between the left and right wheel cylinders was completed according to the pressure distribution formula.

[0039] Based on four typical operating conditions, disturbance / fault conditions such as brake fluid temperature change (-20℃, 80℃), minor pipeline leakage, and pressure sensor drift were introduced to verify the system's robustness and fault tolerance: When the sensor drifts, a moving average filtering algorithm is used with a filter window length of 10 to correct the measured value; when there is a slight leak in the pipeline, the control gain is increased through the disturbance compensation module of the active disturbance rejection controller, with the leakage compensation gain set to 1.2. A fault in the main channel pressure sensor is simulated. The residual chi-square test algorithm is used, with a sliding window length of 15 and a confidence level of 95%. The fault is detected and channel switching is triggered. The system switches to the backup channel to estimate the pressure based on the motor current and displacement sensor signals and obtain the feedback signal.

[0040] During the experiment, the target wheel cylinder hydraulic pressure, actual wheel cylinder hydraulic pressure, pressure response time, overshoot, braking deceleration and other data under various working conditions were recorded synchronously through the data acquisition system. The control effects of the traditional PID control method and the control method of the present invention were compared to verify the effectiveness of the present invention.

[0041] The test results show that the hydraulic pressure control method of the brake-by-wire system of the present invention controls the wheel cylinder hydraulic pressure tracking error within ±3kPa under four typical braking conditions, with an emergency braking pressure response time ≤80ms and overshoot ≤2%, and a conventional braking pressure response time ≤100ms and overshoot ≤1%. Under disturbances such as brake fluid temperature changes, minor pipeline leaks, and sensor drift, the system can still maintain good pressure control accuracy. When the sensor fails, the channel switching response time is ≤20ms, and there is no pressure fluctuation during the switching process. Compared with the traditional PID control method, the pressure tracking accuracy is improved by more than 40%, the response speed is improved by more than 30%, and the overshoot is reduced by more than 60%, effectively verifying the superiority of this method in terms of control accuracy, operating condition adaptability, robustness, and fault tolerance.

[0042] Therefore, this invention provides a hydraulic pressure control method for a brake-by-wire system. It establishes a hydraulic dynamic model integrating pipeline friction and the nonlinear characteristics of hydraulic valve dead zones, employs a complete Lugre dynamic friction model to describe pipeline friction, and introduces brake fluid temperature, longitudinal acceleration, and tire adhesion coefficient as dynamic inputs. A composite architecture of "feedforward compensation + ADRC feedback adjustment" is designed, utilizing an improved PSO algorithm to optimize ESO gain and nonlinear feedback parameters, dynamically adapting to different braking conditions. Multi-sensor signals and vehicle status parameters are acquired in real time, and the total system disturbance is estimated and compensated through ESO. By combining hardware and software analytical redundancy, fault classification and fault tolerance, and vehicle-wide collaborative pressure distribution strategies, this method addresses the problems of insufficient nonlinear compensation, poor condition adaptability, and imbalance between robustness and real-time performance in existing technologies. This achieves precise, rapid, and stable hydraulic pressure control, significantly improving braking accuracy, safety, and reliability.

[0043] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A hydraulic pressure control method for a brake-by-wire system, characterized in that, Includes the following steps: S1. Establish a hydraulic dynamics model for the brake-by-wire system. This model integrates the nonlinear characteristics of brake line friction and hydraulic valve dead zone, and uses the Lugre dynamic friction model to describe the line friction behavior. It also incorporates brake fluid temperature T and vehicle longitudinal acceleration. Tire adhesion coefficient As dynamic input parameters, pressure transmission equations are constructed; S2. Based on the hydraulic dynamics model, a composite control architecture of feedforward compensation and feedback regulation is designed. The feedforward compensation module generates compensation torque according to the output results of the Lugre dynamic friction model, and the feedback regulation module uses an active disturbance rejection controller (ADRC) to track the target hydraulic pressure. S3. The extended state observer (ESO) gain and nonlinear error feedback control law parameters of the active disturbance rejection controller are optimized by improving the particle swarm optimization (PSO) algorithm. The improved PSO algorithm adopts a dynamic learning factor and nonlinear inertial weight adjustment strategy to obtain the objective function of the improved PSO algorithm. S4. Real-time acquisition of wheel cylinder pressure, brake motor current, displacement sensor signals and vehicle status parameters of the brake-by-wire system; estimation of total system disturbance and real-time compensation through the extended state observer of the active disturbance rejection controller; output of wheel cylinder hydraulic pressure control command.

2. The hydraulic pressure control method for a brake-by-wire system according to claim 1, characterized in that, In S1, the hydraulic dynamics model integrates the nonlinear characteristics of brake line friction and hydraulic valve dead zone, and uses the Lugre dynamic friction model to describe the line friction behavior. Its complete expression is: ； ； ； when hour, ; In the formula, For the frictional torque of the braking pipeline, The average stiffness of the bristles. The viscous damping coefficient is... This is the bristle damping coefficient. This represents the average deformation of the bristles on the contact surface. The relative velocity of the pipeline, The rate of change of the average deformation of the bristles. For Stribeck characteristic function, For Coulomb friction torque, For the maximum static friction torque, Stribeck characteristic velocity, It is a natural constant. This is a minimum speed threshold used to avoid numerical singularities; The constructed pressure transmission equation is: ； In the formula, For the wheel cylinder pressure, For hydraulic valve flow gain, For effective control of the voltage of the hydraulic valve, The total leakage coefficient of the system. For the effective area of ​​the wheel cylinder, The speed of the piston movement in the cylinder. The total volume of the hydraulic system. This is the effective volumetric elastic modulus of the brake fluid.

3. The hydraulic pressure control method for a brake-by-wire system according to claim 2, characterized in that, In S1, the process of establishing the hydraulic dynamics model of the brake-by-wire system also includes establishing a nonlinear model of the valve dead zone by combining the flow-pressure characteristic curve of the brake valve, and characterizing the pressure response characteristics within the dead zone range by a piecewise function, wherein the piecewise function is characterized as follows: ； In the formula, For effective control of the voltage of the hydraulic valve, The original control voltage, This is the dead zone threshold.

4. The hydraulic pressure control method for a brake-by-wire system according to claim 3, characterized in that, In S3, the expression for the dynamic learning factor is: ； ； The expression for the nonlinear inertia weight is as follows: ； In the formula, This represents the maximum value of the individual learning factor. This represents the minimum value of the individual learning factor. This represents the maximum value of the social learning factor. This represents the minimum value of the social learning factor. This represents the current iteration number. The maximum number of iterations, This is the adjustment coefficient for individual learning factors. The social learning factor adjustment coefficient. This represents the maximum value of the inertia weight. This represents the minimum value of the inertia weight. The objective function of the improved particle swarm optimization algorithm is: ； In the formula, For the target hydraulic pressure, This is the actual hydraulic pressure. For pressure response time, For overshoot, The weighting coefficients for pressure tracking accuracy are... The pressure response speed weighting coefficient is used. This is a weighting factor for pressure overshoot, and the sum of these factors must be 1. The weighting factor is dynamically adjusted according to the braking conditions, especially during emergency braking. During normal braking , For integration time; The extended state observer (ESO) gain of the active disturbance rejection controller adopts a nonlinear gain function, and the ESO gain expression is: ； ； ； In the formula, This is a pressure estimate. This is an estimate of the rate of change of pressure. This is the estimated total disturbance. For measurement output, To control the gain, , , , For the observer bandwidth, This is a nonlinear factor used to adjust the feedback nonlinearity of the pressure estimate, with a value ranging from 0.8 to 1. This is a nonlinear factor used to adjust the feedback nonlinearity of the pressure change rate estimate, with a value ranging from 0.5 to 0.

8. This is a nonlinear factor used to adjust the feedback nonlinearity of the total disturbance estimate, with a value ranging from 0.1 to 0.5, and satisfying the following conditions: , For nonlinear error functions: ； In the formula, For observation error, This is the threshold for the linear interval.

5. The hydraulic pressure control method for a brake-by-wire system according to claim 4, characterized in that, It also includes operating condition identification and parameter self-tuning steps: Braking conditions are identified using the K-Nearest Neighbors (KNN) algorithm, including emergency braking, normal deceleration, hill braking, and cornering braking; a preset parameter set is called for different braking conditions, and the key parameters of the hydraulic dynamics model are updated in real time using the Recursive Least Squares (RLS) method, including the pipeline friction coefficient and brake fluid viscosity coefficient. The RLS algorithm iterative formula is: ； ； ； in, For the first Time parameter estimates, For recursive gain, For measurement output, For the regression vector, Let be the error covariance matrix. The forgetting factor has a value range of (0.95, 0.99).

6. The hydraulic pressure control method for a brake-by-wire system according to claim 5, characterized in that, It also includes the analytical redundancy design steps for hardware and software integration: A dual-path pressure acquisition mechanism with a main channel and a backup channel is constructed. The main channel directly collects the wheel cylinder pressure through a pressure sensor. The backup channel is based on the brake motor current. With displacement sensor signal Establish a pressure estimation model: ； in, For cylinder wheel pressure estimation, This refers to the motor current-pressure gain coefficient. This is the piston speed-pressure dynamic coefficient. This is the static coefficient of piston displacement-pressure. The model bias is obtained through offline calibration. The residual chi-square test algorithm is used to detect the sensor status in real time. When the main channel sensor fails, it automatically switches to the backup channel to obtain the pressure feedback signal. The switching process is imperceptible.

7. The hydraulic pressure control method for a brake-by-wire system according to claim 6, characterized in that, The implementation process of the residual chi-square test algorithm includes: calculating the residual sequence between the main channel pressure measurement value and the backup channel pressure estimate value. ； The residual sequence follows a mean of 0 and a variance of . The normal distribution; Construct the residual function and make a judgment; when the residual function value is greater than the degrees of freedom... Confidence level is The critical value of the chi-square distribution, i.e. When a main channel sensor malfunction is detected, channel switching is triggered. The residual function is: ； In the formula, This is the length of the sliding window.

8. The hydraulic pressure control method for a brake-by-wire system according to claim 7, characterized in that, In S4, the vehicle state parameters also include the wheel angle of the steer-by-wire system. yaw rate of the vehicle body Based on the vehicle's overall state parameters, the wheel cylinder hydraulic pressure distribution is performed. The formula for the left and right wheel pressure distribution during cornering braking is as follows: ； ； ； ; In the formula, For the left wheel cylinder pressure, For the right wheel cylinder pressure, The distribution coefficient takes the value (0, 1). This refers to the load transfer amount between the left and right wheels. For the vertical load on the tire, For the overall vehicle quality, It is lateral acceleration. For the height of the center of mass, The wheelbase is the distance between the wheels. For vehicle speed, This refers to the vehicle's longitudinal acceleration.

9. The hydraulic pressure control method for a brake-by-wire system according to claim 8, characterized in that, It also includes fault classification and fault tolerance steps: A preset fault level is defined, including sensor drift fault, actuator partial failure fault, and minor pipeline leakage fault. In the case of sensor drift fault, a moving average filtering algorithm is used to correct the measured value. ； In the formula, This is the length of the filtering window; In the event of a partial actuator failure, the system switches to the backup execution channel; in the event of a minor pipeline leak, the control gain is increased through the disturbance compensation module of the active disturbance rejection controller, and the corrected control law is as follows: ； In the formula, The original control signal, To compensate for leakage gain, This is an estimate of the leakage disturbance.

10. The hydraulic pressure control method for a brake-by-wire system according to claim 9, characterized in that, It also includes an adaptive compensation step for the adhesion coefficient: Based on wheel speed and cylinder pressure signals, and combined with the Magic Tire model, the tire-road adhesion coefficient is estimated. The expression for the lateral force of the Magic Tire model is as follows: In the formula, This refers to the lateral force of the tire. Side slip angle, For stiffness factor, For shape factor, As the peak factor, Curvature factor; The precise coefficient of adhesion of the tire is obtained by recursively solving using the magic tire model. , The upper limit of the wheel cylinder hydraulic pressure is dynamically adjusted. ,when At that time, reduce the upper limit of hydraulic pressure to , This is a preset threshold.

Images