Commercial vehicle integrated brake control and sensor system

Abstract

The invention relates to a commercial vehicle integrated brake control and sensor system. The system comprises a multi-source sensor module used for acquiring sensor signals of operation of the commercial vehicle brake control system. The feature extraction module carries out filtering processing on sensor signals and extracts key feature parameters to construct a feature parameter matrix. In the braking decision execution module, a front and rear axle control unit calculates a front and rear axle braking force distribution coefficient based on the characteristic parameter matrix and generates an intelligent braking instruction; and the braking execution unit dynamically adjusts single-wheel braking force distribution according to the intelligent braking instruction to obtain braking execution parameters. And the evaluation optimization module judges the stability of the braking system according to the braking execution parameters and the operation real-time feedback data, and adjusts the braking force distribution parameters of each wheel to generate an optimized braking control strategy. By adopting the system, the multi-source data fusion precision, the dynamic adaptability of braking force distribution and the communication stability under complex working conditions in the braking process of the commercial vehicle can be improved, and the safety and reliability of the braking system are enhanced.

Description

Technical Field

[0001] The present invention belongs to the technical field of vehicle engineering, and in particular relates to an integrated braking control and sensor system for a commercial vehicle. Background Art

[0002] In the field of vehicle engineering technology, the safety and reliability of braking systems are directly related to the lifeblood of transportation. Traditional commercial vehicle braking systems are limited by distributed sensor architectures, fixed-ratio brake force distribution strategies, and non-standardized communication protocols, exposing multiple technical bottlenecks. The independent operation of multiple sensors leads to delayed signal acquisition and weak anti-interference capabilities, making it difficult to accurately reflect vehicle dynamics in real time. The fixed brake force distribution model fails to account for dynamic variables such as load changes and road adhesion, which can easily lead to risks such as brake deviation and tire lock. The general CAN protocol lacks compatibility in a 24V voltage environment, resulting in signal distortion and transmission delays. The lack of a closed-loop optimization mechanism based on braking effect feedback makes it difficult to ensure system stability and energy efficiency, especially in complex conditions such as long downhill slopes and slippery roads. With the growing demand for intelligent and connected commercial vehicles, the development of brake control systems that integrate multi-source data fusion, dynamic brake force distribution, customized communication protocols, and intelligent algorithms has become a key approach to breaking through traditional technical barriers and improving the active safety performance of commercial vehicles. Summary of the Invention

[0003] Based on this, it is necessary to address the above technical issues and provide a commercial vehicle integrated braking control and sensor system that can improve the accuracy of multi-source data fusion, dynamic adaptability of braking force distribution, and communication stability under complex working conditions during the braking process of commercial vehicles.

[0004] In a first aspect, the present application provides an integrated brake control and sensor system for a commercial vehicle, comprising: Multi-source sensor module, used to obtain sensor signals for the operation of commercial vehicle brake control systems.

[0005] The feature extraction module is used to filter the sensor signals to obtain processed sensor data; and extract key feature parameters from each processed sensor data to construct a feature parameter matrix.

[0006] Braking decision execution module, including: The front and rear axle control units are used to calculate the front and rear axle braking force distribution coefficient based on the characteristic parameter matrix combined with the preset braking control model; if the distribution coefficient reaches the preset threshold, the customized 24VCAN chip is used to transmit the distribution coefficient via CAN communication to generate intelligent braking instructions.

[0007] The brake execution unit is used to dynamically adjust the single-wheel braking force distribution according to the intelligent braking instruction to obtain the brake execution parameters.

[0008] The evaluation and optimization module is used to determine the stability of the braking system based on the braking execution parameters combined with real-time feedback data from system operation. If the stability is lower than the preset threshold, the braking force distribution parameters of each wheel are dynamically adjusted based on the model predictive control algorithm to generate an optimized braking control strategy.

[0009] In one embodiment, the feature extraction module includes: The filtering processing unit is used to use a multi-stage adaptive filter to separate the noise of the sensor signal to obtain the noise-reduced signal data.

[0010] Feature calculation unit, used for: Time-frequency domain energy distribution parameters are extracted from the denoised signal data to obtain an initial feature set including amplitude variance and spectrum peaks; the energy distribution parameters include at least one of tire slip rate, deceleration gradient, torque change rate and load transfer rate.

[0011] Perform cross-sensor correlation analysis on the initial feature set to determine the mutual information strength between features and generate a correlation weight matrix.

[0012] The initial feature set is dynamically weighted and fused based on the correlation weight matrix to obtain the feature parameter matrix.

[0013] Matrix optimization unit for: The feature parameter matrix is ​​input into the preset deep autoencoder network for nonlinear mapping, and a low-dimensional embedding vector is output.

[0014] If the reconstruction error of the low-dimensional embedding vector exceeds a preset threshold, the cutoff frequency of the multi-stage adaptive filter and the fusion coefficient of the correlation weight matrix are readjusted to obtain the optimized feature parameter matrix.

[0015] In one embodiment, the front and rear axle control units include: Obtain wheel speed sensor data and vehicle load parameters in the characteristic parameter matrix.

[0016] The dynamic weight factor of the braking control model is determined according to the vehicle load parameters; the dynamic weight factor is used to adjust the calculation weight of the front and rear axle braking force distribution coefficient.

[0017] The wheel speed sensor data are weightedly fused using dynamic weight factors to generate dynamic allocation coefficients.

[0018] Determine whether the dynamic allocation coefficient exceeds a preset allocation coefficient threshold.

[0019] If the allocation coefficient threshold is exceeded, the dynamic allocation coefficient is encoded using the data encapsulation format of 24V CAN communication to generate a transmission data frame with a check code.

[0020] Generate intelligent braking instructions based on the dynamic allocation coefficients in the transmitted data frame.

[0021] In one embodiment, the dynamic allocation coefficient is calculated by the following formula: = ( )

[0022] in, represents the dynamic allocation coefficient, ( ) represents the Sigmoid function, Indicates the number of wheel speed sensors, Indicates the wheel speed sensor data, Indicates the Estimated adhesion coefficient of each wheel, Indicates the The load distribution ratio of each axle, Indicates the The dynamic weight factor of each sensor, represents the slip rate influencing factor, Indicates the deviation between the average wheel speed and the actual vehicle speed. Indicates the reference speed threshold, represents the sensor data fusion function, Represents the load response function.

[0023] In one embodiment, the single-wheel braking force distribution is dynamically adjusted according to the intelligent braking instruction to obtain the braking execution parameters, including: The single wheel slip ratio deviation is calculated based on the wheel speed sensor data; the single wheel slip ratio deviation is obtained by the difference between the current speed and the reference speed.

[0024] The braking force correction value is obtained by using a fuzzy PID control algorithm according to the single wheel slip rate deviation; the fuzzy PID control algorithm dynamically adjusts the PID parameters according to the single wheel slip rate deviation and its change rate.

[0025] Obtain the pressure feedback signal of the brake actuator; the pressure feedback signal contains the current brake line pressure value.

[0026] The target brake pressure gradient is calculated based on the braking force correction amount and the pressure feedback signal.

[0027] The target brake pressure gradient is used to generate brake execution parameters according to a preset period.

[0028] The single-wheel braking force distribution strategy is updated according to the braking execution parameters; the single-wheel braking force distribution strategy includes the weight coefficient of the braking force of each wheel.

[0029] If it is detected that the weight coefficient exceeds a preset threshold, a braking force redistribution instruction is triggered; the braking force redistribution instruction carries a revised set of weight coefficients.

[0030] New braking execution parameters are generated according to the braking force redistribution instruction; the braking execution parameters are transmitted to the corresponding braking execution unit through the bus protocol.

[0031] In one embodiment, the target brake pressure gradient is calculated using the following formula: = ( t ) + + d + (1 + ) in, represents the target brake pressure gradient, ( t ) represents the slip ratio deviation, Indicates the current slip rate, 、 and represents the PID controller parameters, Indicates the rate of change of the brake pressure feedback signal, represents the pressure feedback gain coefficient, 、 represents the nonlinear correction parameter.

[0032] In one embodiment, the evaluation optimization module includes: Stability assessment unit for: Obtain brake execution parameters and real-time feedback data of system operation; brake execution parameters include brake pressure, wheel speed difference and slip rate.

[0033] The stability deviation index is calculated based on the rate of change of the slip ratio weighted by the brake pressure and the wheel speed difference.

[0034] If the stability deviation index exceeds a preset threshold, the model predictive control algorithm is called to generate predictive control parameters; the predictive control parameters include the braking force distribution gradient and the dynamic torque margin.

[0035] A policy generation unit, configured to: The braking force distribution parameters of each wheel are adjusted according to the dynamic torque margin to generate an optimized braking force distribution matrix.

[0036] The braking force distribution matrix is ​​input into the brake controller, which outputs the brake valve current command and hydraulic adjustment command; the hydraulic adjustment command returns the actual pressure feedback value.

[0037] According to the difference between the actual pressure feedback value and the braking force distribution matrix, the constraints of the model predictive control algorithm are updated to generate an optimized braking control strategy.

[0038] In one embodiment, based on the difference between the actual pressure feedback value and the braking force distribution matrix, the constraints of the model predictive control algorithm are updated to generate an optimized braking control strategy, including: Obtain the wheel cylinder pressure gradient and wheel speed fluctuation rate from the braking execution parameters; obtain the tire pressure change and slip rate increment from the real-time feedback data of the system operation.

[0039] The wheel cylinder pressure gradient, wheel speed fluctuation rate and tire pressure change are input into a pre-established deep neural network stability evaluation model to output the current braking stability index.

[0040] If the current braking stability index is lower than the preset threshold, the slip rate increment is extracted as the constraint boundary value of the model predictive control algorithm.

[0041] The multi-objective optimization function is reconstructed according to the boundary values ​​of the constraint conditions to generate a braking force distribution parameter set including feedforward compensation coefficients and dynamic distribution weights.

[0042] The braking force distribution parameter set is processed by embedded control to obtain independent pressure correction values ​​for the four wheels.

[0043] The state parameters of the brake actuator are updated based on the independent pressure correction amount, and the response delay data and hydraulic pulsation characteristics of the actuator are synchronously collected.

[0044] The update frequency of the dynamic allocation weight is adjusted according to the response delay data, and the amplitude threshold of the feedforward compensation coefficient is corrected in combination with the hydraulic pulsation characteristics.

[0045] The constraint boundary conditions of the model predictive control algorithm are iteratively updated based on the revised dynamic allocation weights and feedforward compensation coefficients.

[0046] According to the coupling relationship between the constraint boundary conditions and the vehicle dynamic parameters, a braking control strategy including an execution parameter sequence is obtained.

[0047] In a second aspect, the present application further provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the above system when executing the computer program.

[0048] In a third aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, which implements the above system when executed by a processor.

[0049] The aforementioned integrated braking control and sensor system for commercial vehicles, along with computer equipment and storage media, comprises a multi-source sensor module, a feature extraction module, a braking decision execution module, and an evaluation and optimization module. The multi-source sensor module acquires sensor signals from the commercial vehicle's braking control system. The feature extraction module filters the sensor signals to obtain processed sensor data, extracts key feature parameters, and constructs a feature parameter matrix. In the braking decision execution module, the front and rear axle control units calculate the front and rear axle braking force distribution coefficients based on the feature parameter matrix and a preset braking control model. When the distribution coefficients reach a preset threshold, they generate intelligent braking commands using customized 24V CAN communication data. The braking execution unit dynamically adjusts the individual wheel braking force distribution based on the intelligent braking commands to obtain braking execution parameters. The evaluation and optimization module determines braking system stability based on the braking execution parameters and real-time feedback from system operation. If the stability falls below a preset threshold, it dynamically adjusts the individual wheel braking force distribution parameters based on a model predictive control algorithm to generate an optimized braking control strategy. This system achieves comprehensive braking state awareness through multi-source sensor integration. Feature extraction and intelligent algorithms enable dynamic braking force distribution and communication protocol customization. Furthermore, closed-loop optimization of the braking strategy is achieved through model predictive control. It can improve the accuracy of multi-source data fusion during commercial vehicle braking, the dynamic adaptability of braking force distribution, and communication stability under complex working conditions, enhance the safety and reliability of the braking system under different loads and road conditions, and provide technical support for active safety control of commercial vehicles. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following briefly introduces the drawings required for use in the embodiments or related technical descriptions. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0051] Figure 1 A structural block diagram of an integrated braking control and sensor system for commercial vehicles provided by an embodiment of the present invention; Figure 2 A schematic diagram of a 24V controller area network communication dedicated circuit for a commercial vehicle brake control system provided by an embodiment of the present invention; Figure 3 A schematic diagram of a multi-source sensor signal acquisition and processing circuit for an electronic stability control system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0052] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0053] In one embodiment, Figure 1 As shown, the present application provides an integrated braking control and sensor system for commercial vehicles, which may include: The multi-source sensor module 101 is used to obtain sensor signals of the commercial vehicle brake control system operation.

[0054] Specifically, the multi-source sensor module integrates multiple devices, including a rotor speed sensor, pressure sensor, temperature sensor, integrated angle and torque sensor, triaxial angular velocity sensor, azimuth sensor, lateral acceleration sensor, load sensor, and tire pressure sensor. This module collects critical data such as wheel speed, brake line pressure, brake temperature, driver's operating intention, vehicle posture dynamics, and lateral load changes. The module utilizes hardware redundancy to establish dual acquisition channels for key sensors, and uses isolation circuits to suppress electromagnetic interference. It also supports the CAN bus communication protocol and achieves high-speed, simultaneous acquisition of multi-source data at a 100Hz sampling rate, ensuring real-time and accurate signals.

[0055] The feature extraction module 102 is used to filter the sensor signals to obtain processed sensor data; and extract key feature parameters from each processed sensor data to construct a feature parameter matrix.

[0056] Specifically, the feature extraction module first uses the Kalman filter algorithm to reduce the noise of the sensor's original signal. Through iterative calculations of the system state equation and the measurement equation, it dynamically estimates the true value of the signal, effectively filters out high-frequency noise and random interference, and retains the dynamic change characteristics of the signal (such as sudden changes in speed during sudden braking). The filtered signal is processed to extract key characteristic parameters, including rotor speed, angle change rate, torque value, lateral acceleration, deceleration gradient, and tire slip rate (calculated by the formula ,in, Indicates vehicle speed, represents the wheel angular velocity, represents the wheel radius), and uses Fast Fourier Transform (FFT) to analyze the speed signal in the time and frequency domains, extracting the fundamental frequency and harmonic components to identify abnormal brake vibration characteristics. Ultimately, these characteristic parameters are constructed into a time-varying feature matrix, whose dimensions are dynamically adjusted based on the number of sensors and parameter types.

[0057] The braking decision execution module 103 includes: The front and rear axle control units 1031 are used to calculate the front and rear axle braking force distribution coefficient based on the characteristic parameter matrix combined with the preset braking control model; if the distribution coefficient reaches the preset threshold, the distribution coefficient is transmitted via CAN communication using a customized 24V CAN chip to generate an intelligent braking command.

[0058] Furthermore, the front and rear axle control units calculate the front and rear axle braking force distribution coefficients based on the characteristic parameter matrix and the preset braking control model (supporting both adaptive PID control and fuzzy control modes). In conventional braking scenarios, the adaptive PID control algorithm is used to achieve precise regulation of the braking force by adjusting the proportional (P), integral (I), and differential (D) parameters in real time. In complex working conditions such as ice and snow, slippery roads, or emergency braking, the fuzzy control mode is automatically switched to. The tire slip rate and yaw rate deviation are used as input variables. After fuzzification, rule inference, and defuzzification, the front and rear axle braking force distribution coefficients are output (for example, the front axle braking force calculation formula is ,in, Indicates the vehicle mass, / Indicates the wheelbase parameter, represents the height of the center of mass, Deceleration, The fuzzy control output distribution coefficient is represented by the fuzzy control output. When the distribution coefficient reaches a preset threshold (such as deceleration exceeding a safety threshold or a single-wheel slip ratio greater than 30%), the electronic braking system (EBS) logic is triggered. Using a custom 24V CAN chip and CAN communication (based on the NCV7390 chip with integrated CRC checksum and priority queue mechanism), data is encoded and transmitted to generate intelligent braking commands, ensuring transmission latency of less than 10ms.

[0059] The braking execution unit 1032 is used to dynamically adjust the single-wheel braking force distribution according to the intelligent braking instruction to obtain the braking execution parameters.

[0060] Specifically, after receiving the intelligent braking command, the brake actuator unit activates the angle valve via the communicating angle valve control unit. Combined with dynamic load data from the lateral acceleration sensor, it precisely adjusts the brake chamber pressure, achieving intelligent distribution of braking force per wheel. The unit incorporates a built-in valve spool travel sensor (accuracy ±0.1mm) and a pressure sensor, providing real-time feedback on the angle valve opening and brake pressure, establishing a closed-loop control circuit for brake pressure. Simultaneously, an integral algorithm is used to estimate the braking distance based on the current deceleration and vehicle speed. Feedback data, including the angle valve operating status, pressure value, and estimated braking distance, is encoded and transmitted back to the main control system.

[0061] The evaluation and optimization module 104 is used to determine the stability of the braking system based on the braking execution parameters combined with the real-time feedback data of the system operation; if the stability is lower than the preset threshold, the braking force distribution parameters of each wheel are dynamically adjusted based on the model predictive control algorithm to generate an optimized braking control strategy.

[0062] Specifically, the evaluation and optimization module dynamically assesses braking system stability based on a state-space model and multi-parameter indicators (such as a yaw rate deviation threshold of ±5° / s and an axle load transfer rate threshold of ±15%). A stability deviation index is calculated by weighted integration of brake pressure, wheel speed difference, and slip rate change. If the index exceeds a preset threshold, a model predictive control (MPC) algorithm is invoked for optimization. MPC uses deceleration error, yaw rate error, and control variable change over multiple future time steps as objective functions. This objective function is solved through rolling optimization to dynamically adjust the braking force distribution parameters for each wheel. Further data, such as wheel cylinder pressure gradient, wheel speed fluctuation, and tire pressure change, are collected and fed into a deep neural network stability evaluation model. The MPC constraints are iteratively updated based on the output stability indicators, generating an optimized braking control strategy consisting of feedforward compensation coefficients and dynamic allocation weights. The system supports fault tolerance mechanisms (such as switching interpolation algorithms in the event of sensor failure). The hardware design complies with the ISO 26262 ASIL-D functional safety standard and includes an over-the-air (OTA) remote upgrade interface for online algorithm and model optimization.

[0063] The aforementioned integrated braking control and sensor system for commercial vehicles includes a multi-source sensor module, a feature extraction module, a braking decision execution module, and an evaluation and optimization module. The multi-source sensor module acquires sensor signals from the commercial vehicle's braking control system. The feature extraction module filters the sensor signals to obtain processed sensor data, extracts key feature parameters, and constructs a feature parameter matrix. In the braking decision execution module, the front and rear axle control units calculate the front and rear axle braking force distribution coefficients based on the feature parameter matrix and a preset braking control model. When the distribution coefficients reach a preset threshold, a customized 24V CAN chip transmits data via CAN communication to generate intelligent braking commands. The braking execution unit dynamically adjusts the individual wheel braking force distribution based on the intelligent braking commands to obtain braking execution parameters. The evaluation and optimization module determines braking system stability based on the braking execution parameters and real-time feedback from system operation. If the stability falls below a preset threshold, the braking force distribution parameters for each wheel are dynamically adjusted based on a model predictive control algorithm to generate an optimized braking control strategy. This system achieves comprehensive braking state awareness through multi-source sensor integration. Feature extraction and intelligent algorithms enable dynamic braking force distribution and communication protocol customization. Combined with model predictive control, closed-loop optimization of the braking strategy is achieved. It can improve the accuracy of multi-source data fusion during commercial vehicle braking, the dynamic adaptability of braking force distribution, and communication stability under complex working conditions, enhance the safety and reliability of the braking system under different loads and road conditions, and provide technical support for active safety control of commercial vehicles.

[0064] In one embodiment, the feature extraction module 102 may include: The filtering processing unit 1021 is used to perform noise separation on the sensor signal using a multi-stage adaptive filter to obtain noise-reduced signal data.

[0065] The feature calculation unit 1022 is configured to: Step S101: extracting time-frequency domain energy distribution parameters from the denoised signal data to obtain an initial feature set including amplitude variance and spectrum peaks; the energy distribution parameters include at least one of tire slip rate, deceleration gradient, torque change rate, and load transfer rate.

[0066] Step S102 : performing cross-sensor correlation analysis on the initial feature set, determining the mutual information strength between features, and generating a correlation weight matrix.

[0067] Step S103 : performing dynamic weighted fusion on the initial feature set based on the correlation weight matrix to obtain a feature parameter matrix.

[0068] The matrix optimization unit 1023 is used to: Step S104: input the feature parameter matrix into a preset deep autoencoder network for nonlinear mapping, and output a low-dimensional embedding vector.

[0069] Preferably, the deep autoencoder network comprises an input layer, hidden layers, and an output layer, and is trained layer by layer to achieve nonlinear compression and reconstruction of the feature parameter matrix. The network optimizes weight parameters by minimizing reconstruction error (e.g., mean squared error). When the output low-dimensional embedding vector cannot accurately restore the original matrix (i.e., the reconstruction error exceeds a preset threshold), a parameter adjustment mechanism is triggered. Firstly, the cutoff frequency of the multi-stage adaptive filter is dynamically adjusted based on the characteristic frequency distribution to suppress residual noise. Secondly, the fusion coefficient of the correlation weight matrix is ​​recalculated based on the change in mutual information strength to strengthen the weight of key features.

[0070] Step S105 : If the reconstruction error of the low-dimensional embedding vector exceeds a preset threshold, the cutoff frequency of the multi-stage adaptive filter and the fusion coefficient of the correlation weight matrix are readjusted to obtain an optimized feature parameter matrix.

[0071] Specifically, the filtering processing unit uses a multi-stage adaptive filter to remove noise from the sensor signals, obtaining de-noised signal data. The feature calculation unit extracts time-frequency energy distribution parameters from the de-noised data, forming an initial feature set consisting of amplitude variance and spectral peaks (parameters covering tire slip, deceleration gradient, torque change rate, and load transfer rate). This initial feature set is then subjected to cross-sensor correlation analysis, calculating the strength of mutual information between features to generate a correlation weight matrix. Based on this matrix, the initial feature set is dynamically weighted and fused to construct a feature parameter matrix. The matrix optimization unit inputs the feature parameter matrix into a deep autoencoder network for nonlinear mapping, outputting a low-dimensional embedding vector. If the reconstruction error of the embedding vector exceeds a threshold, the filter cutoff frequency and weight matrix fusion coefficients are readjusted to generate an optimized feature parameter matrix.

[0072] This embodiment effectively improves the noise reduction accuracy and feature relevance of multi-source signals through multi-stage adaptive filtering and cross-sensor correlation analysis. A dynamic weighted fusion mechanism ensures that feature parameters are aligned with real-time operating conditions. The nonlinear dimensionality reduction capabilities of the deep autoencoder network exploit the data's latent features, and a reconstructed error feedback mechanism achieves closed-loop optimization of the feature extraction process. This module significantly enhances the reliability and feature representation of sensor data, providing high-dimensional, low-redundancy, high-quality data support for subsequent braking decisions and strengthening the system's dynamic adaptability to complex operating conditions.

[0073] In one embodiment, the front and rear axle control units may include the following steps: Step S201: Obtain wheel speed sensor data and vehicle load parameters in a characteristic parameter matrix.

[0074] Step S202 , determining a dynamic weight factor of the braking control model according to the vehicle load parameter; the dynamic weight factor is used to adjust the calculation weight of the front and rear axle braking force distribution coefficient.

[0075] Preferably, determining the dynamic weight factor of the braking control model may include the following steps: first, analyzing the relationship between the vehicle load parameters and the braking force requirements of the front and rear axles. Generally speaking, the greater the vehicle load, the greater the braking force requirement of the corresponding bridge. Then, by establishing a mathematical model or based on empirical data, determine the weight adjustment rules for the front and rear axle braking force distribution coefficients under different load conditions. For example, the dynamic weight factor can be calculated based on the proportional relationship between the rated load and the actual load of the vehicle, as well as the load distribution of the front and rear axles. The specific calculation process may involve the calibration and adjustment of some parameters to ensure that the dynamic weight factor can accurately reflect the impact of the vehicle load on the braking force distribution. The determined dynamic weight factor can make the braking control model more adaptable to different working conditions and improve the performance and safety of the braking system.

[0076] Step S203 : performing weighted fusion on the wheel speed sensor data using the dynamic weight factor to generate a dynamic allocation coefficient.

[0077] Step S204 , determining whether the dynamic allocation coefficient exceeds a preset allocation coefficient threshold.

[0078] Step S205: If the allocation coefficient threshold is exceeded, the dynamic allocation coefficient is encoded using the data encapsulation format of 24V CAN communication to generate a transmission data frame with a check code.

[0079] Step S206: Generate an intelligent braking instruction according to the dynamic allocation coefficient in the transmission data frame.

[0080] Specifically, the system first obtains wheel speed sensor data and vehicle load parameters from the characteristic parameter matrix. Based on the vehicle load parameters, the dynamic weighting factor of the braking control model is determined. This factor is used to adjust the calculation weights for the front and rear axle braking force distribution coefficients. Next, the dynamic weighting factor is used to perform a weighted fusion on the wheel speed sensor data to generate a dynamic distribution coefficient. A determination is then made as to whether the dynamic distribution coefficient exceeds a preset distribution coefficient threshold. If so, the dynamic distribution coefficient is encoded using the 24V CAN chip and the CAN communication data encapsulation format, generating a transmission data frame with a checksum. Finally, an intelligent braking command is generated based on the dynamic distribution coefficient in the transmission data frame.

[0081] This embodiment can comprehensively consider the actual operating status and load conditions of the vehicle by acquiring wheel speed sensor data and vehicle load parameters. The introduction of dynamic weighting factors makes the calculation of the front and rear axle braking force distribution coefficients more flexible and accurate, and can be dynamically adjusted according to different load conditions. The weighted fusion of wheel speed sensor data generates a dynamic distribution coefficient, further improving the rationality of the braking force distribution. Determining the dynamic distribution coefficient and performing corresponding processing ensures that intelligent braking commands can be generated in a timely manner when necessary, ensuring the safety and effectiveness of the braking system. The entire process realizes the refined management of braking control and improves the performance and reliability of the commercial vehicle braking system.

[0082] In one embodiment, the dynamic allocation coefficient is calculated by the following formula: = ( )

[0083] in, represents the dynamic allocation coefficient, ( ) represents the Sigmoid function, Indicates the number of wheel speed sensors, Indicates the wheel speed sensor data, Indicates the Estimated adhesion coefficient of each wheel, Indicates the The load distribution ratio of each axle, Indicates the The dynamic weight factor of each sensor, represents the slip rate influencing factor, Indicates the deviation between the average wheel speed and the actual vehicle speed. Indicates the reference speed threshold, represents the sensor data fusion function, Represents the load response function.

[0084] Preferably, The dynamic distribution coefficient is used to quantify the distribution ratio of the braking force between the front and rear axles. The larger the value, the higher the proportion of the front axle braking force.

[0085] No. The estimated adhesion coefficient of each wheel is dynamically estimated using the extended Kalman filter algorithm combined with data such as wheel speed and lateral acceleration. The value range is [0,1], which represents the grip of the road surface (e.g., it approaches 0 on icy and snowy roads and approaches 1 on dry roads).

[0086] The deviation between the average wheel speed and the actual vehicle speed reflects the wheel slip trend. = ,in, Indicates the reference speed threshold, which is used to normalize the speed deviation. It is usually the vehicle's current speed or a preset safe speed. Indicates the actual vehicle speed.

[0087] Sensor data fusion function, used to calculate the braking demand of a single wheel by combining wheel speed and adhesion coefficient (such as = , represents the road condition correction factor).

[0088] Load response function, usually a nonlinear function (such as exponential function = ), used to characterize the dynamic impact of load changes on braking force distribution, represents the response slope, Indicates the load balancing point.

[0089] In one embodiment, dynamically adjusting the single-wheel braking force distribution according to the intelligent braking instruction to obtain the braking execution parameter may include the following steps: Step S301 : Calculate the single wheel slip ratio deviation based on wheel speed sensor data; the single wheel slip ratio deviation is obtained by the difference between the current speed and the reference speed.

[0090] Step S302 : A braking force correction value is obtained by using a fuzzy PID control algorithm according to the single wheel slip ratio deviation; the fuzzy PID control algorithm dynamically adjusts PID parameters according to the single wheel slip ratio deviation and its change rate.

[0091] Step S303: Obtain a pressure feedback signal from the brake actuator; the pressure feedback signal includes a current brake line pressure value.

[0092] Step S304 : Calculate the target braking pressure gradient according to the braking force correction value and the pressure feedback signal.

[0093] Step S305 : generating a braking execution parameter based on the target braking pressure gradient according to a preset period.

[0094] Step S306: updating the single-wheel braking force distribution strategy according to the braking execution parameters; the single-wheel braking force distribution strategy includes a weight coefficient of the braking force of each wheel.

[0095] Step S307: If it is detected that the weight coefficient exceeds a preset threshold, a braking force redistribution instruction is triggered; the braking force redistribution instruction carries a revised weight coefficient set.

[0096] Step S308 : generating new braking execution parameters according to the braking force redistribution instruction; and transmitting the braking execution parameters to the corresponding braking execution unit via the bus protocol.

[0097] Specifically, based on wheel speed sensor data, the slip ratio deviation of a single wheel is calculated by taking the difference between the current speed and a reference speed. Based on this deviation, a fuzzy PID control algorithm is employed to dynamically adjust the PID parameters, taking into account the deviation and its rate of change, to determine a braking force correction. Simultaneously, the brake actuator's pressure feedback signal (including the current brake line pressure) is obtained and combined with the braking force correction to calculate a target brake pressure gradient. This target brake pressure gradient is then used to generate braking actuation parameters at a preset interval, which are used to update the single-wheel braking force distribution strategy (including the braking force weighting coefficients for each wheel). If the weighting coefficients exceed a preset threshold, a braking force redistribution command is triggered, generating new braking actuation parameters with the corrected weighting coefficients and transmitting them to the corresponding brake actuation unit via the bus protocol.

[0098] This embodiment achieves adaptive output of braking force corrections through precise calculation of single-wheel slip deviations and dynamic adjustment of a fuzzy PID control algorithm. This, combined with pressure feedback signals, forms a closed-loop control system, enhancing the accuracy and responsiveness of brake pressure regulation. Braking execution parameters are generated and the distribution strategy is updated at a preset periodic interval to ensure that braking force distribution matches real-time operating conditions. Exceeding the weight coefficient limit triggers a redistribution mechanism, promptly correcting abnormal braking force distribution and avoiding the risk of single-wheel overload or locking. This closed-loop process of "deviation detection - algorithm correction - pressure control - strategy update" significantly improves the stability and safety of commercial vehicle braking systems, particularly under complex road conditions or with uneven loads.

[0099] In one embodiment, the target brake pressure gradient may be calculated using the following formula: = ( t ) + + d + (1 + ) in, represents the target brake pressure gradient, ( t ) represents the slip ratio deviation, Indicates the current slip rate, 、 and represents the PID controller parameters, Indicates the rate of change of the brake pressure feedback signal, represents the pressure feedback gain coefficient, 、 represents the nonlinear correction parameter.

[0100] Preferably, ( t ) The slip rate deviation is the difference between the current slip rate and the ideal slip rate, which is used to measure the degree of deviation between the wheel slip state and the expected state.

[0101] The current slip rate, calculated using wheel speed sensor data, indicates the degree of wheel slip during braking and is a key indicator for evaluating braking effectiveness and safety.

[0102] 、 and Indicates the PID controller parameters, namely the proportional coefficient, integral coefficient and differential coefficient. Determines the speed and strength of the controller's response to the deviation; integral coefficient Used to eliminate the steady-state error of the system; differential coefficient The changing trend of the deviation can be predicted and adjustments can be made in advance to improve the dynamic performance of the system.

[0103] This embodiment comprehensively considers multiple factors, including slip ratio deviation, current slip ratio, PID controller parameters, brake pressure feedback signal change rate, and nonlinear correction parameters. By precisely calculating the target brake pressure gradient, it achieves refined control of the braking system. By real-time monitoring and adjusting the brake pressure gradient, the wheel slip ratio is maintained within a reasonable range, effectively preventing wheel lock and improving vehicle stability and controllability during braking, thereby enhancing braking safety. Dynamic adjustment of PID controller parameters and the introduction of nonlinear correction parameters enable the control algorithm to adaptively adjust to varying braking conditions and vehicle states, optimizing braking effectiveness, reducing braking distance, and improving braking efficiency. Considering the brake pressure feedback signal change rate and pressure feedback gain coefficient allows the system to better adapt to the dynamic characteristics of the brake actuator and the actual operating environment, improving system reliability and stability.

[0104] In one embodiment, the evaluation and optimization module 104 may include: The stability evaluation unit 1041 is configured to: Step S401, obtaining braking execution parameters and real-time feedback data of system operation; the braking execution parameters include braking pressure, wheel speed difference and slip ratio.

[0105] Step S402 : Calculate a stability deviation index based on the rate of change of the slip ratio obtained by weighted fusion of the braking pressure and the wheel speed difference.

[0106] Step S403: If the stability deviation index exceeds a preset threshold, the model predictive control algorithm is called to generate predictive control parameters; the predictive control parameters include the braking force distribution gradient and the dynamic torque margin.

[0107] The policy generating unit 1042 is configured to: Step S404 : adjusting the braking force distribution parameters of each wheel according to the dynamic torque margin to generate an optimized braking force distribution matrix.

[0108] Furthermore, adjusting the braking force distribution parameters of each wheel according to the dynamic torque margin includes the following steps. First, it is necessary to determine the relationship between the dynamic torque margin and the braking force of each wheel. This can be achieved by establishing a mathematical model or based on empirical data. Generally speaking, the larger the dynamic torque margin, the more appropriate the braking force of the corresponding wheel can be increased to fully utilize the braking capacity of the vehicle. Then, based on the determined relationship, the braking force distribution parameters of each wheel are adjusted. The specific adjustment method can be linear or nonlinear, depending on the actual situation and control requirements. For example, the braking force distribution weight of each wheel can be adjusted according to a certain proportional coefficient based on the size of the dynamic torque margin.

[0109] Step S405 , inputting the braking force distribution matrix into the brake controller, outputting a brake valve current instruction and a hydraulic adjustment instruction; the hydraulic adjustment instruction returns an actual pressure feedback value.

[0110] Step S406 : updating the constraints of the model predictive control algorithm according to the difference between the actual pressure feedback value and the braking force distribution matrix, and generating an optimized braking control strategy.

[0111] Specifically, the stability assessment unit acquires braking parameters (including brake pressure, wheel speed difference, and slip ratio) as well as real-time feedback from system operation. It then calculates a stability deviation index based on the brake pressure and wheel speed difference, weighted by the slip ratio's rate of change. If the stability deviation index exceeds a preset threshold, the model predictive control algorithm is invoked to generate predictive control parameters, including the braking force distribution gradient and dynamic torque margin. The strategy generation unit then adjusts the braking force distribution parameters for each wheel based on the dynamic torque margin, generating an optimized braking force distribution matrix. This matrix is ​​input into the brake controller, which then outputs brake valve current commands and hydraulic pressure adjustment commands. The hydraulic pressure adjustment commands also return actual pressure feedback values. Finally, based on the difference between the actual pressure feedback value and the braking force distribution matrix, the constraints of the model predictive control algorithm are updated, generating an optimized braking control strategy.

[0112] Through the collaborative work of the stability assessment unit and the strategy generation unit, this embodiment enables the system to monitor the operating status of the braking system in real time and promptly adjust the control strategy based on the stability deviation index. This effectively improves the stability and safety of the braking system and reduces unstable factors during braking, such as wheel locking and brake deviation. Furthermore, by optimizing the braking force distribution matrix, it can improve braking efficiency, shorten braking distance, and enhance the vehicle's braking performance. Furthermore, by updating the constraints of the model predictive control algorithm based on actual pressure feedback values, the control strategy can be more adaptable to changes in actual operating conditions, enhancing the system's adaptability and robustness.

[0113] In one embodiment, updating the constraints of the model predictive control algorithm based on the difference between the actual pressure feedback value and the braking force distribution matrix to generate an optimized braking control strategy may include the following steps: Step S501 , obtaining the wheel cylinder pressure gradient and wheel speed fluctuation rate from the braking execution parameters; obtaining the tire pressure change and slip rate increment from the real-time feedback data of the system operation.

[0114] Step S502: Input the wheel cylinder pressure gradient, wheel speed fluctuation rate, and tire pressure change into a pre-established deep neural network stability evaluation model to output the current braking stability index.

[0115] The deep neural network stability assessment model preferably utilizes a multi-layer, fully connected architecture. The input layer receives normalized multidimensional features such as wheel cylinder pressure gradient, wheel speed fluctuation, and tire pressure change. The hidden layer extracts nonlinear feature associations using the Reluctant Unified Unit (ReLU) activation function. The output layer is a single node (a stability index ranging from 0 to 1, with lower values ​​indicating poorer stability). The model is trained using labeled real-world braking data (covering dry, wet, and icy conditions), using the mean squared error (MSE) as the loss function. Weight parameters are iteratively updated using the Adam optimizer, ultimately achieving a stability assessment accuracy of ≥95% on the test set. This model automatically captures the complex coupling between brake pressure, wheel speed fluctuation, and tire pressure changes. Compared to traditional linear assessment methods, it improves the accuracy of identifying nonlinear conditions (such as sudden local load changes caused by abnormal tire pressure) by over 30%, providing a more reliable basis for subsequent MPC algorithms.

[0116] Step S503 : If the current braking stability index is lower than a preset threshold, the slip ratio increment is extracted as a constraint boundary value of the model predictive control algorithm.

[0117] Step S504 : reconstructing the multi-objective optimization function according to the boundary values ​​of the constraint conditions to generate a braking force distribution parameter set including a feedforward compensation coefficient and a dynamic distribution weight.

[0118] Step S505 : performing embedded control processing on the braking force distribution parameter set to obtain independent pressure correction values ​​for the four wheels.

[0119] Step S506 : updating the state parameters of the brake actuator based on the independent pressure correction amount, and synchronously collecting the response delay data and hydraulic pulsation characteristics of the actuator.

[0120] Step S507 : adjusting the update frequency of the dynamic allocation weight according to the response delay data, and correcting the amplitude threshold of the feedforward compensation coefficient in combination with the hydraulic pulsation characteristics.

[0121] Step S508 , iteratively updating the constraint boundary conditions of the model predictive control algorithm based on the revised dynamic allocation weights and feedforward compensation coefficients.

[0122] Step S509 : obtaining a braking control strategy including an execution parameter sequence according to the coupling relationship between the constraint boundary conditions and the vehicle dynamics parameters.

[0123] Specifically, the system acquires brake actuation parameters such as wheel cylinder pressure gradients and wheel speed fluctuations, as well as tire pressure changes and slip ratio increments from real-time feedback data. The wheel cylinder pressure gradients, wheel speed fluctuations, and tire pressure changes are input into a pre-established deep neural network stability assessment model, which outputs a current braking stability index. If this index falls below a preset threshold, the slip ratio increment is extracted as a constraint boundary value for the model predictive control (MPC) algorithm. Based on the boundary values, a multi-objective optimization function is reconstructed to generate a braking force distribution parameter set consisting of feedforward compensation coefficients and dynamic distribution weights. The parameter set is processed through embedded control to obtain independent pressure corrections for the four wheels. Based on these corrections, the brake actuator state parameters are updated, while response delay data and hydraulic pulsation characteristics are simultaneously collected. The update frequency of the dynamic distribution weights is adjusted based on the response delay, and the amplitude threshold of the feedforward compensation coefficients is modified based on the hydraulic pulsation characteristics. The MPC algorithm's constraint boundary conditions are iteratively updated based on the corrected parameters. Finally, a braking control strategy containing a sequence of execution parameters is generated based on the coupling relationship between the constraints and vehicle dynamics parameters.

[0124] This embodiment integrates multi-dimensional data such as wheel cylinder pressure gradient and wheel speed fluctuation rate, and uses deep neural networks to achieve accurate assessment of braking stability, avoiding the limitations of single parameter judgment. Using the slip rate increment as an MPC constraint condition can directly control the potential risk of instability, improving the targeted optimization target. The independent pressure correction value generated by embedded processing enables fine-grained adjustment of the braking force of a single wheel; the dynamic correction mechanism combined with the actuator response delay and hydraulic pulsation characteristics makes up for the shortcomings of traditional control in responding to hardware characteristics. It significantly improves the dynamic adaptability of the braking strategy to complex working conditions, enhances system stability and control accuracy, and effectively reduces the risk of skidding and tailspin in scenarios such as long downhill slopes and sudden load changes.

[0125] In one embodiment, Figure 2 As shown, the present application provides a 24V controller area network communication dedicated circuit for a commercial vehicle brake control system, which may include: 1. Main control chip area Core Component: NCV7390 chip (CAN bus transceiver), responsible for CAN signal level conversion and driving, supports 24V input voltage, and is compatible with the ISO 11898 standard. Peripheral Components: R34 (150Ω): Current-limiting resistor on the chip power supply, protecting the power circuit. C67 (100nF / 50V): Power supply filter capacitor, suppresses high-frequency noise and stabilizes the VCC power supply.

[0126] 2. Signal transmission and matching area Differential signal interface: CANH (pin 13) and CANL (pin 11): Connected to the bus via 162kΩ resistors R9 and 22kΩ resistors R12. A 220pF capacitor C63 is connected in parallel between CANH and CANL to filter high-frequency interference. 88Ω resistor R18 is a termination resistor to eliminate signal reflections. Inductor component: ACT45B-510-2P-TL003. The primary (pins 1-2) connects to the CHBH / CHBL pins of the NCV7390. The secondary (pins 3-4) outputs the AS-CANH / AS-CANL signals, achieving electrical isolation and preventing ground loop interference.

[0127] 3. Protection and filtering area ESD protection: An ESD4 (62R electrostatic protection device) connected in parallel between CANH / CANL and ground suppresses transient overvoltages, complying with the ISO 10605 electrostatic protection standard. Signal filtering: C137 (47nF / 50V): Connected in series with the CANH signal path to filter out common-mode interference; C118 (220pF / 50V): Connected in parallel with the power supply to further reduce ripple.

[0128] 4. Control and drive area Enable control: ENCHB (pin 10) and ENCLB (pin 2) are pulled up to VCC via a 10kΩ resistor (R15). They receive the MCU-CAN2_ENCHB / ENCLB control signal, activating the chip's transceiver function. Data transmission and reception: CAN_TX2 (pin 4) and CAN_RX2 (pin 7) connect to the MCU's CAN controller to enable data transmission and reception. The 288Ω resistor (R38) is a signal current-limiting resistor.

[0129] This embodiment realizes high-speed differential transmission and electrical isolation of data, ensuring the stability of communication under complex working conditions.

[0130] In one embodiment, Figure 3As shown, the present application provides a multi-source sensor signal acquisition and processing circuit for an electronic stability control system, which may include: 1. Main control chip area Core component: The SCC3134 chip (ESC sensor interface chip) features a built-in high-precision analog-to-digital converter and supports SPI protocol communication, allowing access to sensor signals such as wheel speed, pressure, and acceleration. Power management: VCC (B) and 3.3VCC: External 100nF capacitor C119 and 1μF capacitors C123 / C124 are used for high-frequency and low-frequency filtering, respectively, to ensure a stable 3.3V power supply. VDDA / VDDD: Analog / digital power supply isolation, separated by a GND network to reduce mutual interference.

[0131] 2.SPI communication interface area Data transmission pins: MOSI (Master Out Slave In), MISO (Master In Slave Out), SCK (Clock), and CSB (Chip Select): Connect to the MCU (28C2757) via the SPI bus. 1.91kΩ resistor R69 acts as a pull-up resistor for the MOSI signal to ensure data transmission reliability. Reserved pins (such as TVEXTRESN and EXTRESYRS0) are reserved for extended functionality and can be used to connect external reference voltages or backup sensor signals.

[0132] 3. Sensor signal access area Signal conditioning circuit: Wheel speed sensor signal: Accessed through the BP interface, processed by the internal amplification circuit (in conjunction with a peripheral resistor network, not marked in the figure), and converted into a digital value; Pressure / acceleration signal: Accessed through an unmarked analog input pin. The chip integrates a programmable gain amplifier (PGA) that can dynamically adjust the amplification factor according to the signal range.

[0133] 4. Grounding and isolation areas Ground network: GNDA (analog ground), GND (digital ground), and GNDD (power ground) are independently partitioned and connected through 0Ω resistors or ferrite beads to reduce ground loop interference; capacitors C123 / C124 (1μF / 50V) are connected across the power supply and ground to enhance the power supply filtering effect.

[0134] 5. Accessibility Area Reset circuit: The unlabeled RESET pin (implied in the chip function in the figure) is connected to an external RC reset circuit (simplified in the figure) to ensure chip initialization when powered on; reserved port: NO (B010480-01000C) is the hardware version identifier, used for production debugging and compatibility management.

[0135] This embodiment converts the sensor analog signal into a digital quantity and transmits it to the main control unit via the SPI protocol.

[0136] It should be understood that, although the various steps in the flowcharts involved in the various embodiments described above are displayed in sequence according to the instructions of the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowcharts involved in the various embodiments described above can include multiple steps or multiple stages, and these steps or stages are not necessarily executed and completed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of steps or stages in other steps.

[0137] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the aforementioned integrated braking control and sensor system for commercial vehicles when executing the computer program.

[0138] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0139] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to the partial description of the method embodiments. The device embodiments described above are merely illustrative, wherein the components described as separate parts may or may not be physically separated, and the parts displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the disclosed solution. A person of ordinary skill in the art can understand and implement it without expending creative work.

[0140] The above-described embodiments merely represent several implementation methods of the embodiments of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the patent application. It should be noted that a person skilled in the art may make various modifications and improvements without departing from the concept of the embodiments of the present application, and these modifications and improvements fall within the scope of protection of the embodiments of the present application.

Claims

1. A commercial vehicle integrated brake control and sensor system, characterized in that: The system comprises: Multi-source sensor module, used to obtain sensor signals for the operation of commercial vehicle brake control systems; A feature extraction module is used to filter the sensor signals to obtain processed sensor data; and extract key feature parameters from each of the processed sensor data to construct a feature parameter matrix; Braking decision execution module, including: The front and rear axle control units are configured to calculate a front and rear axle braking force distribution coefficient based on the characteristic parameter matrix and a preset braking control model; if the distribution coefficient reaches a preset threshold, the distribution coefficient is transmitted via CAN communication using a customized 24V CAN chip to generate an intelligent braking command; A brake execution unit, configured to dynamically adjust the single-wheel braking force distribution according to the intelligent braking instruction to obtain a brake execution parameter; An evaluation and optimization module is used to determine the stability of the braking system based on the braking execution parameters combined with real-time feedback data from system operation; if the stability is lower than a preset threshold, the braking force distribution parameters of each wheel are dynamically adjusted based on the model predictive control algorithm to generate an optimized braking control strategy.

2. The system according to claim 1, wherein: The feature extraction module includes: A filtering processing unit, configured to perform noise separation on the sensor signal using a multi-stage adaptive filter to obtain noise-reduced signal data; Feature calculation unit, used for: Extracting time-frequency domain energy distribution parameters from the noise-reduced signal data to obtain an initial feature set including amplitude variance and spectrum peaks; the energy distribution parameters include at least one of tire slip rate, deceleration gradient, torque change rate, and load transfer rate; Performing a cross-sensor correlation analysis on the initial feature set to determine the mutual information strength between features and generate a correlation weight matrix; Performing dynamic weighted fusion on the initial feature set based on the correlation weight matrix to obtain a feature parameter matrix; Matrix optimization unit for: Input the feature parameter matrix into a preset deep autoencoder network for nonlinear mapping, and output a low-dimensional embedding vector; If the reconstruction error of the low-dimensional embedding vector exceeds a preset threshold, the cutoff frequency of the multi-stage adaptive filter and the fusion coefficient of the correlation weight matrix are readjusted to obtain an optimized feature parameter matrix.

3. The system according to claim 1, wherein: The front and rear axle control units include: Obtaining wheel speed sensor data and vehicle load parameters in the characteristic parameter matrix; Determining a dynamic weight factor of a braking control model according to the vehicle load parameter; the dynamic weight factor is used to adjust the calculation weight of the front and rear axle braking force distribution coefficient; Performing weighted fusion on the wheel speed sensor data using the dynamic weight factor to generate a dynamic allocation coefficient; Determining whether the dynamic allocation coefficient exceeds a preset allocation coefficient threshold; If the allocation coefficient threshold is exceeded, the dynamic allocation coefficient is encoded using the data encapsulation format of the 24VCAN communication to generate a transmission data frame with a check code; An intelligent braking instruction is generated according to the dynamic allocation coefficient in the transmission data frame.

4. The system according to claim 3, characterized in that The dynamic allocation coefficient is calculated by the following formula: = ( ) in, represents the dynamic allocation coefficient, ( ) represents the Sigmoid function, Indicates the number of wheel speed sensors, Indicates the wheel speed sensor data, Indicates the Estimated adhesion coefficient of each wheel, Indicates the The load distribution ratio of each axle, Indicates the The dynamic weight factor of each sensor, represents the slip rate influencing factor, Indicates the deviation between the average wheel speed and the actual vehicle speed. Indicates the reference speed threshold, represents the sensor data fusion function, Represents the load response function.

5. The system according to claim 1, wherein: The dynamically adjusting the single-wheel braking force distribution according to the intelligent braking instruction to obtain the braking execution parameter includes: Calculating a single wheel slip ratio deviation based on wheel speed sensor data; the single wheel slip ratio deviation is obtained by the difference between the current speed and the reference speed; A fuzzy PID control algorithm is used to obtain a braking force correction value according to the single wheel slip rate deviation; the fuzzy PID control algorithm dynamically adjusts PID parameters according to the single wheel slip rate deviation and its change rate; Obtaining a pressure feedback signal from the brake actuator; the pressure feedback signal includes a current brake line pressure value; calculating a target braking pressure gradient according to the braking force correction amount and the pressure feedback signal; generating a braking execution parameter according to a preset period using the target braking pressure gradient; updating a single-wheel braking force distribution strategy according to the braking execution parameter; wherein the single-wheel braking force distribution strategy includes a weight coefficient of each wheel braking force; If it is detected that the weight coefficient exceeds a preset threshold, a braking force redistribution instruction is triggered; the braking force redistribution instruction carries the modified weight coefficient set; New braking execution parameters are generated according to the braking force redistribution instruction; and the braking execution parameters are transmitted to the corresponding braking execution unit through the bus protocol.

6. The system according to claim 5, characterized in that The target brake pressure gradient is calculated using the following formula: = ( t ) + + d + (1 + ) in, represents the target brake pressure gradient, ( t ) represents the slip ratio deviation, Indicates the current slip rate, 、 and represents the PID controller parameters, Indicates the rate of change of the brake pressure feedback signal, represents the pressure feedback gain coefficient, 、 represents the nonlinear correction parameter.

7. The system according to claim 1, wherein: The evaluation and optimization module includes: Stability assessment unit for: Obtaining the braking execution parameters and real-time feedback data of the system operation; the braking execution parameters include braking pressure, wheel speed difference and slip ratio; A stability deviation index is calculated by weighting the braking pressure and the wheel speed difference and fusing the slip rate with the slip rate; If the stability deviation index exceeds a preset threshold, a model predictive control algorithm is called to generate predictive control parameters; the predictive control parameters include a braking force distribution gradient and a dynamic torque margin; A policy generation unit, configured to: Adjusting the braking force distribution parameters of each wheel according to the dynamic torque margin to generate an optimized braking force distribution matrix; The braking force distribution matrix is ​​input into the brake controller, which outputs a brake valve current instruction and a hydraulic adjustment instruction; the hydraulic adjustment instruction returns an actual pressure feedback value; According to the difference between the actual pressure feedback value and the braking force distribution matrix, the constraint conditions of the model predictive control algorithm are updated to generate an optimized braking control strategy.

8. The system according to claim 7, characterized in that The updating of the constraint conditions of the model predictive control algorithm according to the difference between the actual pressure feedback value and the braking force distribution matrix to generate an optimized braking control strategy includes: Obtaining the wheel cylinder pressure gradient and wheel speed fluctuation rate from the braking execution parameters; obtaining the tire pressure change and slip rate increment from the real-time feedback data of the system operation; Inputting the wheel cylinder pressure gradient, wheel speed fluctuation rate, and tire pressure change into a pre-established deep neural network stability evaluation model to output a current braking stability index; If the current braking stability index is lower than a preset threshold, extracting the slip ratio increment as a constraint boundary value of the model predictive control algorithm; Reconstructing a multi-objective optimization function according to the boundary values ​​of the constraint conditions to generate a braking force distribution parameter set including a feedforward compensation coefficient and a dynamic distribution weight; Performing embedded control processing on the braking force distribution parameter set to obtain independent pressure correction values ​​for the four wheels; updating the state parameters of the brake actuator based on the independent pressure correction amount, and synchronously collecting the response delay data and hydraulic pulsation characteristics of the actuator; adjusting the update frequency of the dynamic allocation weight according to the response delay data, and correcting the amplitude threshold of the feedforward compensation coefficient in combination with the hydraulic pulsation characteristics; Iteratively updating the constraint boundary conditions of the model predictive control algorithm based on the revised dynamic allocation weights and feedforward compensation coefficients; According to the coupling relationship between the constraint boundary conditions and the vehicle dynamics parameters, a braking control strategy including an execution parameter sequence is obtained.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the system according to any one of claims 1 to 8 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method of any one of claims 1 to 8 are implemented.

Images