Intelligent electro-hydraulic composite braking system of electric vehicle and control method thereof

Abstract

The invention relates to the technical field of electric vehicle braking, and discloses an intelligent electro-hydraulic composite braking system of an electric vehicle and a control method of the intelligent electro-hydraulic composite braking system. The system comprises a braking state sensing module, a risk assessment module, a strategy collaborative generation module and a control updating module. The braking state sensing module collects vehicle running state data and road surface environment information in real time through a multi-source sensor, and a braking working condition feature set and environment feature information are constructed. The braking risk assessment module obtains a comprehensive braking risk value in combination with the environment characteristic information, the initial braking parameters of the vehicle, the current braking request and the state information of the electro-hydraulic actuator; when the risk value is abnormal, a braking strategy collaborative generation module generates a plurality of electro-hydraulic composite braking distribution schemes, calculates a collaborative execution interference degree and an intervention evaluation value, screens an optimal scheme and generates an optimized braking control strategy; and the brake control updating module updates brake control parameters in real time according to an optimization strategy, generates brake operation log information and adapts to complex working conditions.

Description

Technical Field

[0001] This invention relates to the field of electric vehicle braking technology, specifically to an intelligent electro-hydraulic hybrid braking system for electric vehicles and its control method. Background Technology

[0002] With the rapid development of the electric vehicle industry, the performance requirements of vehicle braking systems are constantly increasing. They must not only meet basic deceleration and stopping functions, but also consider energy recovery efficiency, driving stability, and adaptability to complex road conditions. Traditional pure hydraulic braking systems rely on mechanical transmission structures, which have limitations in response speed and struggle to dynamically adjust the braking force distribution ratio according to real-time vehicle conditions and road environment. Especially in complex conditions such as slippery roads and emergency avoidance maneuvers, uneven braking force distribution can easily occur, affecting vehicle braking safety.

[0003] While existing electro-hydraulic hybrid braking systems incorporate electronic control units (ECUs) for braking decisions, most systems rely on a single sensor to collect vehicle operating data, such as wheel speed sensors or acceleration sensors. This fails to comprehensively acquire crucial information such as the vehicle's longitudinal and lateral dynamic states and road surface adhesion coefficients, resulting in a one-sided braking condition feature set. In the braking risk assessment phase, current technologies often only consider the vehicle's current speed and brake pedal travel, neglecting the real-time status of the electro-hydraulic actuators, such as solenoid valve response delays and hydraulic line pressure fluctuations. This leads to significant discrepancies between the calculated comprehensive braking risk value and actual operating conditions, making it difficult to accurately predict potential risks such as wheel lock-up and skidding during braking.

[0004] In terms of braking strategy generation, traditional systems often employ preset fixed braking distribution schemes, lacking adaptability to dynamic environments. When the vehicle's driving environment changes abruptly, such as transitioning from a dry road to a flooded one, the preset scheme cannot adjust the electro-hydraulic braking ratio in time, easily causing interference in braking coordination and leading to decreased vehicle stability. Furthermore, while some systems can generate multiple braking distribution schemes, they lack a scientific intervention evaluation value calculation model, relying solely on simple comparisons of braking distances to select schemes. This ignores key factors affecting stability during braking, such as changes in vehicle attitude and tire contact pressure distribution, making it difficult for the final braking control strategy to achieve an optimal balance between braking efficiency and driving stability.

[0005] Existing braking control update modules mostly employ periodic update mechanisms, which cannot dynamically adjust the update frequency according to real-time changes in braking conditions. In scenarios with extremely high response speed requirements, such as emergency braking, control parameter updates are prone to lag, affecting braking performance. Furthermore, most systems do not fully record the braking operation process, hindering subsequent tracing and analysis of braking faults, further reducing system reliability and maintenance convenience. Summary of the Invention

[0006] The purpose of this invention is to provide an intelligent electro-hydraulic hybrid braking system for electric vehicles and its control method, so as to solve the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides an intelligent electro-hydraulic hybrid braking system for electric vehicles, the system comprising: The braking state perception module collects vehicle operating status data and road environment information in real time through multi-source sensors to construct a braking condition feature set and braking environment feature information. The braking risk assessment module, based on braking environment characteristic information, combined with the vehicle's initial braking parameters, current braking request, and current time electro-hydraulic actuator status information, obtains a comprehensive braking risk value based on a braking condition feature set. The braking strategy collaborative generation module generates multiple electro-hydraulic composite braking allocation schemes based on the braking condition feature set when the comprehensive braking risk value is abnormal. Combining the braking environment feature information, it calculates the collaborative execution interference degree corresponding to the current time when the electro-hydraulic actuator state switches to the reference state, obtains the intervention evaluation value of each braking allocation scheme on vehicle stability, selects the best electro-hydraulic composite braking allocation scheme, and generates the optimized braking control strategy for the current time. The braking control update module updates the braking control parameters in real time according to the optimized braking control strategy at the current time and generates braking operation log information.

[0008] Preferably, the multi-source sensor includes a wheel speed sensor, a pressure sensor, an inertial measurement unit, and a road surface recognition sensor; Each element in the braking condition feature set corresponds to a braking condition feature information, which includes tire slip ratio, braking force distribution coefficient and road surface adhesion coefficient extracted based on sensor data. The braking environment characteristics information includes road surface unevenness and humidity levels in different areas.

[0009] Preferably, the initial braking parameters of the vehicle include a reference electro-hydraulic braking ratio and a first pressure tolerance; The current time electro-hydraulic actuator status information includes the real-time electro-hydraulic braking ratio and the second pressure tolerance; The first pressure tolerance represents the maximum value of the difference between the rated pressure and the actual pressure of the brake line under reference conditions; The second pressure tolerance represents the maximum value of the difference between the rated pressure and the actual pressure of the brake line under the current condition.

[0010] Preferably, when the braking risk assessment module calculates the comprehensive braking risk value, it records the comprehensive braking risk value of the vehicle based on the braking condition feature set as the braking risk index. The braking risk index is obtained by a weighted combination of the first pressure tolerance, the second pressure tolerance, and the deviation of ideal braking force from the current braking request and braking condition characteristics. When the actual braking force is higher than the ideal braking force, the deviation is taken as a positive value; otherwise, it is taken as a negative value.

[0011] Preferably, when the braking strategy collaborative generation module generates multiple electro-hydraulic composite braking distribution schemes, it extracts the safe braking force boundary nodes corresponding to each adhesion coefficient in the braking condition feature set, marks the safe boundaries corresponding to each adhesion coefficient in the braking force distribution interval, and records the feasible domain between the current braking force request value and each safe boundary node as the braking distribution safe domain. The brake allocation safety domain includes one or more brake allocation subdomains; The gradient descent algorithm is used to solve the electro-hydraulic hybrid braking allocation scheme corresponding to each braking allocation subdomain.

[0012] Preferably, when the braking strategy coordination generation module calculates the coordination execution interference degree, it obtains the command response sequence and pressure regulation sequence during the state switching process of the electro-hydraulic actuator; The degree of interference in coordinated execution is obtained by normalized weighting of instruction response overlap and pressure regulation amplitude. The instruction response overlap is calculated based on the ratio of the total number of instructions required for state switching to the actual number of reusable instructions. The pressure regulation amplitude is taken as the maximum value of the second pressure tolerance during the state switching process.

[0013] Preferably, when the braking strategy collaborative generation module calculates the intervention evaluation value of each braking distribution scheme on vehicle stability, a state switching execution flag is introduced; The intervention evaluation value is obtained by the ratio of the comprehensive braking risk value, the degree of interference in coordinated execution, and the weighted product of the path length and road surface smoothness corresponding to the braking distribution scheme. When the state transition execution flag is true, the cooperative execution interference degree is included in the calculation; otherwise, the baseline interference degree is used instead.

[0014] Preferably, when the braking control update module updates the braking control parameters, it corrects the output commands of the hydraulic adjustment unit and the motor braking unit according to the electro-hydraulic distribution ratio in the optimized braking control strategy. The braking operation log information includes the strategy generation time, the braking allocation scheme adopted, the braking condition feature set, the braking environment feature information, and the electro-hydraulic actuator status information.

[0015] Preferably, the system further includes a braking performance verification module, which inverts the braking process based on the braking operation log information, calculates the degree of agreement between the actual braking distance and the theoretical braking distance, and generates a braking performance evaluation report. When the consistency is lower than the threshold, the braking performance verification module triggers a system recalibration command. The recalibration command is sent to the braking state perception module and the braking risk assessment module to update the feature extraction algorithm and risk calculation parameters.

[0016] Preferably, the present invention also includes an intelligent electro-hydraulic hybrid braking control method for electric vehicles, applied to an intelligent electro-hydraulic hybrid braking system for electric vehicles as described above, the method comprising the following steps: By collecting vehicle operating status data and road environment information in real time through multi-source sensors, a braking condition feature set and braking environment feature information are constructed. Based on the braking environment feature information, combined with the vehicle's initial braking parameters, current braking request, and current time electro-hydraulic actuator status information, a comprehensive braking risk value based on the braking condition feature set is calculated. When the comprehensive braking risk value is abnormal, multiple electro-hydraulic composite braking distribution schemes are generated based on the braking condition feature set. The cooperative execution interference degree corresponding to the switching of the electro-hydraulic actuator state to the baseline state at the current time is calculated in combination with the braking environment feature information. Then, the intervention evaluation value of each braking distribution scheme on vehicle stability is obtained, and the best electro-hydraulic composite braking distribution scheme is selected to generate the optimized braking control strategy at the current time. The braking control parameters are updated in real time according to the optimized braking control strategy, and braking operation log information is generated.

[0017] Compared with the prior art, the beneficial effects of the present invention are: By using a multi-source sensor to collect data through the braking state perception module, comprehensive vehicle operating status data and road environment information can be obtained. Compared with the single sensor acquisition method, a more complete and accurate braking condition feature set and braking environment feature information can be constructed, making the system's understanding of the current braking condition more comprehensive. This provides more reliable basic information for subsequent braking risk assessment and strategy generation, avoiding braking decision deviations caused by incomplete information collection.

[0018] During the assessment process, the braking risk assessment module fully integrates braking environment characteristics, vehicle initial braking parameters, current braking requests, and electro-hydraulic actuator status information. This breaks through the limitations of traditional assessment methods that only focus on a few parameters, and can more comprehensively consider various factors affecting braking risk. As a result, it obtains a comprehensive braking risk value that is more in line with actual operating conditions. It can more accurately identify potential risk situations during braking, providing a reasonable basis for subsequent braking strategy adjustments, and enabling the braking system to remain highly sensitive to risks under different operating conditions.

[0019] When the comprehensive braking risk value is abnormal, the braking strategy collaborative generation module generates multiple electro-hydraulic composite braking distribution schemes based on the braking condition feature set. This breaks free from the constraints of traditional fixed schemes and provides the system with more options to adapt to different risk scenarios. Simultaneously, by calculating the collaborative execution interference degree when the electro-hydraulic actuator switches to the baseline state, the intervention evaluation value of each scheme on vehicle stability is obtained. This allows for a scientific evaluation of different schemes from the perspective of vehicle stability, rather than relying solely on single indicators such as braking distance. The selected optimal electro-hydraulic composite braking distribution scheme better balances braking efficiency and vehicle stability, effectively avoiding abnormal vehicle posture problems caused by unreasonable braking distribution. Even in complex scenarios such as sudden changes in road conditions, reasonable braking distribution can ensure vehicle driving stability.

[0020] The braking control update module updates braking control parameters in real time based on the optimized braking control strategy. Compared to traditional periodic updates, this allows for a more timely response to changes in braking conditions, ensuring that braking control parameters always match current braking requirements and preventing braking performance issues caused by delayed parameter updates. Simultaneously, it generates braking operation logs, comprehensively recording various key data during the braking process, including changes in braking control parameters, electro-hydraulic actuator operating status, and vehicle operation data. This provides detailed information support for subsequent troubleshooting and maintenance of the braking system, facilitating technicians in tracing abnormal situations during braking, improving system maintenance convenience and reliability, and reducing system downtime caused by difficulties in troubleshooting. Attached Figure Description

[0021] Figure 1 This is a timing diagram of the intelligent electro-hydraulic hybrid braking system for electric vehicles described in this invention; Figure 2 A flowchart defining the initial braking parameters and actuator status information for the vehicle; Figure 3 A diagram illustrating the multi-stage collaborative optimization analysis of a vehicle's intelligent electro-hydraulic hybrid braking system. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] Please see Figure 1 The present invention provides an intelligent electro-hydraulic hybrid braking system for electric vehicles, the system comprising a braking state perception module, a braking risk assessment module, a braking strategy collaborative generation module, and a braking control update module.

[0024] The braking state perception module collects vehicle operating status data and road environment information in real time through multi-source sensors, constructing a braking condition feature set and braking environment feature information. The braking risk assessment module, based on the braking environment feature information, combined with the vehicle's initial braking parameters, current braking request, and current electro-hydraulic actuator status information, obtains a comprehensive braking risk value based on the braking condition feature set. When the comprehensive braking risk value is abnormal, the braking strategy collaborative generation module generates multiple electro-hydraulic composite braking allocation schemes based on the braking condition feature set; combined with the braking environment feature information, it calculates the collaborative execution interference degree corresponding to the current time when the electro-hydraulic actuator state switches to the baseline state, obtains the intervention evaluation value of each braking allocation scheme on vehicle stability, selects the optimal electro-hydraulic composite braking allocation scheme, and generates an optimized braking control strategy for the current time. The braking control update module updates the braking control parameters in real time according to the optimized braking control strategy for the current time and generates braking operation log information. This system achieves intelligent collaborative control of electro-hydraulic composite braking through modular design, improving the vehicle's braking response accuracy and stability.

[0025] Example 1: See Figure 2The multi-source sensors include wheel speed sensors, pressure sensors, inertial measurement units, and road surface recognition sensors. The wheel speed sensors are installed at the wheel hubs of the four wheels of the vehicle, monitoring the speed changes of each wheel in real time and transmitting the data to the braking status sensing module via the CAN bus. The pressure sensors are embedded in the hydraulic lines between the brake master cylinder and the wheel cylinders, capturing brake fluid pressure fluctuations with a sampling period of 20ms. The inertial measurement unit is fixed at the center of gravity of the vehicle chassis, collecting triaxial acceleration and angular velocity data. The road surface recognition sensor is integrated into the front bumper of the vehicle, using a fusion scheme of millimeter-wave radar and camera to scan the texture features of the road surface ahead. The braking condition feature set is stored in the vehicle memory in matrix form. Each element in the matrix corresponds to a braking condition feature. The tire slip ratio is calculated from the wheel speed sensor data and the vehicle's longitudinal speed. The vehicle's longitudinal speed is generated by integrating the acceleration data output by the inertial measurement unit through Kalman filtering. The braking force distribution coefficient is derived from the differences in the readings of the pressure sensors of each wheel cylinder and the wheel speed change rate, and is specifically expressed as the ratio of the braking force between the front and rear axles. The road surface adhesion coefficient is dynamically matched by fusing the material analysis results output by the road surface recognition sensor with the real-time tire slip ratio curve features through a preset μ-Slip relationship mapping table.

[0026] The braking environment characteristic information is constructed into a hierarchical data structure. Road surface unevenness data comes from the scanning results of road surface undulation height by millimeter-wave radar. Road areas are divided at 0.1m intervals, and the unevenness level of each area is marked. Humidity level is cross-validated by rain sensors and ambient temperature and humidity sensors. When the rain sensor detects precipitation, it activates the camera to analyze the road surface reflectivity and comprehensively determines the humidity level as low / medium / high. The vehicle's initial braking parameters are loaded during system power-on initialization. The baseline electro-hydraulic braking ratio is set to the default value of 40% electric motor braking and 60% hydraulic braking. The first pressure tolerance is defined as the maximum allowable range of the difference between the upper limit of the rated pipeline pressure allowed by the braking system under baseline conditions and the current actual pressure. This value is preset by the vehicle manufacturer to 2.5MPa according to the braking system specifications. The current electro-hydraulic actuator status information is refreshed every 200ms. The real-time electro-hydraulic braking ratio is obtained by converting the torque output percentage fed back by the motor brake controller with the pressure feedback value of the hydraulic regulating unit. The second pressure tolerance is dynamically calculated based on the real-time monitoring data of the pressure sensor, taking the maximum difference between the rated pressure and the actual pressure in each wheel cylinder pipeline at the current moment. When the vehicle is braking in a curve, the data of the outer wheel is calculated first.

[0027] The braking state perception module runs in a real-time operating system environment. Wheel speed sensor data and inertial measurement unit data are first timestamped, and feature change trends within the most recent 500ms are extracted using a sliding window mechanism. The raw point cloud data output by the road surface recognition sensor is rasterized to generate a 10cm×10cm resolution road surface characteristic distribution map, which is then combined with GPS positioning information to construct a road segment feature index. Tire slip ratio calculation adopts an incremental update strategy. When a sudden change in wheel speed is detected exceeding a threshold, an emergency sampling mode is triggered, continuously collecting data from three wheels at a 5ms interval for transient slip ratio analysis. The braking force distribution coefficient calculation incorporates a temperature compensation factor. When the pressure sensor detects that the brake fluid temperature exceeds 80℃, a high-temperature correction coefficient is automatically activated. The road surface adhesion coefficient matching process employs a two-level verification mechanism. The primary matching is based on the position of the current slip ratio in the μ-Slip curve. The secondary verification judges the matching consistency by comparing the slip ratio differences of adjacent wheels. When the slip ratio deviation between the left and right wheels exceeds 15%, a road surface asymmetry adhesion alarm is activated. The road surface unevenness in the braking environment feature information is smoothed using a cubic spline interpolation algorithm. Pre-scan data is loaded 200ms in advance when the vehicle is about to enter a new road section. A hysteresis buffer mechanism is set for humidity level determination. After the rain sensor stops detecting precipitation, the high humidity level is maintained for 120 seconds to avoid frequent switching. The vehicle's initial braking parameters are stored in a tamper-proof memory. When the system detects a brake line replacement, the reference electro-hydraulic braking ratio recalibration process is automatically triggered. The first pressure tolerance is set differently for the vehicle under full load and no load conditions, and the configuration table is dynamically switched through the air suspension height sensor data. The current time electro-hydraulic actuator status information acquisition process is monitored by a hardware watchdog. When the data fluctuation exceeds the threshold within three consecutive refresh cycles, the system switches to the backup sensor channel. The real-time electro-hydraulic braking ratio calculation integrates the torque inverted from the motor phase current and the torque converted from the hydraulic pressure, and takes the weighted average of the two as the final output value. The second pressure tolerance monitoring is set with a pressure gradient limit. When the pressure change rate exceeds 5MPa / s in a single cycle, the current value is immediately frozen and the system diagnostics are initiated. The braking condition feature set update adopts a dual-mode approach of event-driven and periodic scanning. Under normal conditions, the feature matrix is ​​fully updated every 100ms, and when an emergency braking signal is detected, it switches to a high-speed update mode of 10ms. The feature matrix storage uses a circular buffer structure, retaining historical data from the most recent 20 seconds for trend backtracking. Braking environment feature information is linked with high-precision map data. When the vehicle enters a known map section, pre-stored historical road unevenness records are prioritized, and real-time scanning is only initiated for unmapped sections. Humidity level data is integrated with the meteorological service API, and regional humidity prediction values ​​obtained from the network are used when the vehicle enters enclosed environments such as tunnels. The vehicle's initial braking parameters are updated synchronously during OTA upgrades. The system automatically compares the differences between the old and new parameter tables and generates a change log. The first pressure tolerance setting uses an adaptive learning mechanism to dynamically adjust the upper limit value based on brake disc wear sensor data.The electro-hydraulic actuator status information acquisition is equipped with a multi-verification mechanism to check the range rationality of the real-time electro-hydraulic braking ratio. When the sum of the motor torque and hydraulic output is detected to exceed the vehicle's maximum braking force, the safety fuse is immediately triggered.

[0028] The braking state perception module is equipped with a sensor fault tolerance strategy. When the wheel speed sensor fails, it switches to a vehicle speed estimation mode based on the inertial measurement unit. When the pressure sensor fails, it uses the motor back electromotive force to calculate the braking force. When the road surface recognition sensor malfunctions, it retrieves historical adhesion coefficient data from the cache. A confidence weighting factor is added during the construction of the braking condition feature set, and slip rate data in high-noise environments is automatically downweighted. Braking environment feature information activates a conservative mode under extreme weather conditions, and when hail is detected, the road surface unevenness level is forcibly increased to the highest level. The vehicle's initial braking parameters automatically switch to a dedicated configuration set in trailer mode, with the first pressure tolerance proportionally amplified based on the trailer weight sensor feedback value. The current time electro-hydraulic actuator status information is specially marked when energy recovery is limited, and when the power battery temperature exceeds 45℃, the motor contribution in the real-time electro-hydraulic braking ratio decays according to the temperature curve. During system operation, the timeliness of each feature parameter is continuously monitored, and expired data is automatically marked and a re-acquisition process is triggered.

[0029] Example 2: When calculating the comprehensive braking risk value, the braking risk assessment module establishes a braking risk index. This index integrates the first pressure tolerance, the second pressure tolerance, and the deviation of the current braking request from the ideal braking force. The first pressure tolerance is read from the vehicle's initial braking parameters and serves as the upper limit of the pressure tolerance under the baseline state. The second pressure tolerance is obtained in real time from the electro-hydraulic actuator status information, reflecting the pressure margin of the current braking line. The current braking request is obtained through the brake pedal travel sensor or the automatic driving control command. The ideal braking force is dynamically calculated based on the tire slip ratio and road adhesion coefficient in the braking condition feature set: when the tire slip ratio is within the preset optimal range, the maximum theoretical braking force under that adhesion coefficient is taken; when the slip ratio deviates from the range, the ideal value is adjusted according to the preset attenuation curve. The actual braking force is obtained by converting the wheel cylinder pressure value measured by the pressure sensor. The deviation is defined as the algebraic difference between the actual value and the ideal value. When the actual value is higher than the ideal value, the deviation is positive; otherwise, it is negative. The weighted combination process employs a three-layer weighting structure. The first pressure tolerance weight is fixed at 0.3, the second pressure tolerance weight is linearly adjusted according to vehicle speed, and the deviation weight is negatively correlated with the road adhesion coefficient. The final braking risk index is output through a normalized weighted sum formula, and a risk anomaly flag is triggered when the index value exceeds 0.75.

[0030] The braking strategy collaborative generation module initiates a multi-scheme generation process after the risk anomaly flag is activated. First, it extracts the road surface adhesion coefficients of each wheel from the stored braking condition feature set, and calculates the safe braking force boundary node for each wheel. The safe boundary node is determined by querying a pre-set adhesion coefficient-maximum braking force mapping table, which is pre-calibrated based on the vehicle's full load weight and tire model. A three-dimensional coordinate system is established within the braking force distribution range, with the X-axis representing the front axle braking force distribution ratio, the Y-axis the rear axle braking force distribution ratio, and the Z-axis the electro-hydraulic braking ratio. The safe boundary nodes of each wheel are projected onto the corresponding coordinate plane. After the current braking request value is converted into coordinate points, its spatial position is compared with each safe boundary node, forming a braking distribution safety domain defined by the convex hull algorithm. This safety domain is divided into multiple braking distribution subdomains according to the electro-hydraulic ratio gradient, each subdomain representing a specific ratio combination range of electric motor braking and hydraulic braking. The optimal distribution point for each subdomain is solved based on the gradient descent algorithm. During algorithm initialization, a starting point is set at the center of the subdomain, and the algorithm iteratively searches along the braking force distribution coefficient gradient direction with the objective function of minimizing braking force fluctuations. Each iteration calculates the braking force fluctuation value at the current point. The fluctuation value is obtained by simulating the root mean square difference between the actual braking force of each wheel and the requested value. The search terminates when the improvement is less than 1% after three consecutive iterations, and the electro-hydraulic composite braking distribution scheme corresponding to the subdomain is output. A dynamic correction mechanism is set in the braking risk index calculation process. When the inertial measurement unit detects that the vehicle pitch angle exceeds 3 degrees, the deviation weight is automatically increased by 20%. In low-temperature environments, the second pressure tolerance is compensated for temperature based on the brake fluid viscosity detection value. Load transfer compensation is introduced in the calculation of safety boundary nodes. When the vehicle acceleration exceeds 0.5g, the front and rear axle boundary nodes are recalculated according to the dynamic axle load. The construction of the braking distribution safety domain adopts a real-time collision detection algorithm. When the requested value exceeds the physical feasible range under the current adhesion coefficient, the safety domain boundary is automatically shrunk to the nearest feasible point. The gradient descent algorithm sets up a parallel computing architecture. The solution process of each braking distribution subdomain is distributed to an independent processor core, and the scheme generation of all subdomains is completed within 50ms. The scheme output format includes the motor braking torque value, hydraulic target pressure value, and transition time parameters, forming a complete set of electro-hydraulic composite braking distribution schemes. Data updates to the braking condition feature set trigger a recalculation of risk indicators; a new round of risk assessment is immediately initiated when the change in road surface adhesion coefficient exceeds 0.2. A failure protection strategy is implemented for ideal braking force calculation, switching to an estimation mode based on wheel speed difference when the tire slip ratio sensor malfunctions. Deviation in the weighted combination is subject to amplitude limiting; when the positive deviation exceeds 30% of the maximum braking force, it is forcibly zeroed. The safety boundary node mapping table supports online updates, automatically loading the characteristic parameters of the new tire when the system detects a tire replacement. Historical data is incorporated into the construction process of the braking distribution safety domain; when a vehicle repeatedly enters a similar road segment, the previously valid safety domain shape template is invoked. The iteration step size of the gradient descent algorithm is adaptively adjusted, automatically reducing the step size on low-adhesion surfaces to improve search accuracy.

[0031] After the electro-hydraulic hybrid braking distribution scheme is generated, a stability pre-verification is performed. The yaw rate variation under each scheme is simulated using a vehicle dynamics model, eliminating schemes that result in a yaw rate exceeding 8 degrees / second. The scheme set is stored using a priority queue structure, sorted in ascending order of braking force fluctuation value for subsequent filtering. When the system is in a cornering braking condition, asymmetric braking force distribution schemes are generated first, increasing the upper limit of braking force on the outer wheels by 15%. Under regenerative braking limited conditions, the scheme generation process automatically increases the weight ratio of hydraulic braking to ensure that the total braking force meets the requested requirements. Each distribution scheme is marked with an effective timestamp and validity period; schemes not adopted after the timeout are automatically discarded. The system monitors the feasibility of each scheme in real time. When a hydraulic unit pressure response delay exceeds 100ms, the execution risk level is marked in the scheme.

[0032] The braking risk assessment module and the strategy generation module establish a two-way communication mechanism. When a newly generated braking allocation scheme reduces the risk index below the threshold, the subsequent scheme generation process is immediately terminated. Detailed logs are recorded during the risk index calculation process, including the basis for each weight and intermediate calculation results for post-analysis. During system initialization, a module self-check is performed, verifying that the risk index output range meets the design expectation of 0-1 by injecting test data. The construction results of the braking allocation safety domain are visualized and output to the debugging interface, showing the positional relationship between the safety boundary and the requested value in a 3D mesh diagram. The convergence process of the gradient descent algorithm records the number of iterations and the final fluctuation value. If convergence is not achieved after more than 50 iterations, the algorithm automatically switches to a heuristic search algorithm. All generated schemes are packaged into data packets and transmitted to the intervention evaluation value calculation unit, completing the preliminary work for the collaborative generation of braking strategies.

[0033] For example, an electric vehicle equipped with an intelligent electro-hydraulic hybrid braking system is traveling on the Beijing-Hong Kong-Macau Expressway at a speed of 110 km / h. Suddenly, the millimeter-wave radar of the braking status perception module detects a stationary obstacle 150 meters ahead, and the system immediately triggers a braking risk assessment process. The wheel speed sensors show that all four wheels are rotating at 1024 rpm, the pressure sensors report a current hydraulic pressure of 2.1 MPa, and the inertial measurement unit collects a longitudinal acceleration of -0.02g. The road surface recognition sensor identifies the characteristics of the wet asphalt road surface through laser scanning, and the braking condition feature set is updated in real time: left front wheel slip ratio 0.18, right front wheel slip ratio 0.20, braking force distribution coefficient 65% front axle / 35% rear axle, and road surface adhesion coefficient 0.52. The braking environment characteristic information is simultaneously labeled with a road unevenness level of 2 and a humidity level of 3 (heavy rain condition).

[0034] The braking risk assessment module calls the vehicle's initial parameters: a baseline electro-hydraulic braking ratio of 40:60 and a first pressure tolerance of 2.8 MPa. The current electro-hydraulic actuator status shows a real-time braking ratio of 35% motor / 65% hydraulic, with a calculated second pressure tolerance of 1.9 MPa (rated pressure 4.0 MPa minus actual pressure 2.1 MPa). The current braking request is generated by the autonomous driving system, and the required braking force is converted to a hydraulic pressure of 3.8 MPa. The ideal braking force is calculated based on an adhesion coefficient of 0.52 and an optimal slip ratio of 0.2, corresponding to a hydraulic pressure of 3.2 MPa. The actual pressure of 2.1 MPa is lower than the ideal value, with a negative deviation of -1.1 MPa. The comprehensive braking risk value calculation uses dynamic weights: a first pressure tolerance weight of 0.3 (fixed), a second pressure tolerance weight of 0.4 (increased by 0.1 due to vehicle speed > 100 km / h), and a deviation weight of 0.3 (decreased by 0.1 due to high humidity). Substituting these values ​​into the formula yields a risk index of 0.72, exceeding the threshold of 0.7 triggers a risk anomaly flag. After the braking strategy collaborative generation module is activated, the adhesion coefficients of each wheel are extracted from the braking condition feature set: left front 0.50, right front 0.54, left rear 0.51, and right rear 0.53. A preset mapping table is queried to determine the safety boundary nodes of each wheel: the maximum braking force corresponding to the left front wheel is 3.5MPa, the right front wheel is 3.8MPa, the left rear wheel is 2.9MPa, and the right rear wheel is 3.1MPa. A coordinate system is established in the three-dimensional braking force distribution space: X-axis front axle distribution ratio (60%-80%), Y-axis rear axle distribution ratio (20%-40%), and Z-axis electro-hydraulic ratio (30%-50%). The current requested value (front axle 70% / rear axle 30% / motor 40%) is mapped to the spatial point P(70,30,40). The safety domain generated by the convex hull algorithm contains three subdomains: subdomain A (electro-hydraulic ratio 35%-42%), subdomain B (43%-47%), and subdomain C (48%-50%).

[0035] The gradient descent algorithm iteratively searches within subdomain A, starting at point (72, 28, 38), with the objective function set to minimize the fluctuation of braking force across all four wheels. The first iteration calculates a front left wheel pressure of 3.15 MPa with a 72% front axle distribution, which has a 0.35 MPa margin from the wheel's safety boundary of 3.5 MPa. The right rear wheel pressure is calculated at 2.58 MPa with a 28% rear axle distribution, close to the safety boundary of 3.1 MPa. Adjusting the direction along the Z-axis by increasing the motor proportion to 40% reduces the fluctuation value by 12%. After five iterations, the fluctuation value converges to its minimum level at point (70, 30, 41), resulting in scheme A: 41% motor / 59% hydraulic, 70% front axle / 30% rear axle, with a target pressure of 3.8 MPa. The same process applies to output scheme B in subdomain B: motor 45% / hydraulic 55%, front axle 68% / rear axle 32%; output scheme C in subdomain C: motor 49% / hydraulic 51%, front axle 65% / rear axle 35%.

[0036] During the feasibility verification phase, vehicle dynamics model simulations showed that Option C resulted in a peak yaw rate of 7.2 degrees / second under humidity level 3 conditions (exceeding the safety threshold of 6 degrees / second), and this option was eliminated. Options A and B were ultimately added to the candidate queue and transmitted to the intervention evaluation value calculation unit. The entire risk assessment and option generation process took 82 milliseconds, completing strategy preparation when the vehicle approached the obstacle at a distance of 100 meters. The braking control update module executed the preferred option A. The hydraulic adjustment unit received a target pressure command of 3.8 MPa, and the PID controller increased the duty cycle of the inlet valve to 65%. The electric motor braking unit received a torque command of 285 Nm, and the electric motor controller established regenerative braking torque within 50 ms. Wheel speed sensor data showed that the vehicle speed dropped to 60 km / h 3.2 seconds after braking was applied, and the pressure sensor recorded that the hydraulic pressure stabilized within the range of 3.75 ± 0.05 MPa within 150 ms. The braking operation log records the event timestamp 2023-09-12T14:30:05.228. Stored parameters include electro-hydraulic ratio, inter-axle distribution ratio, and target pressure value. Feature set data shows the dynamic change in slip ratio from 0.18 to 0.25. Environmental information records that the humidity level remained at level 3 throughout the braking process. During the execution of the plan, the road surface recognition sensor detected a water accumulation area 20 meters ahead, and the braking condition feature set updated the adhesion coefficient to 0.41 in real time. The braking risk assessment module immediately initiated a recalculation, and the risk index jumped to 0.79. Based on the new feature set, the strategy generation module generated a supplementary plan D within 15ms: motor 38% / hydraulic 62%, front axle 75% / rear axle 25%. This plan was activated and overwrote the original plan the moment the vehicle entered the water accumulation area. The system achieved seamless switching of braking strategies from dry to water accumulation areas through dynamic response, and the vehicle finally came to a complete stop 8.2 meters before the obstacle.

[0037] See Figure 3 The system presents the key parameter coupling relationships and dynamic response characteristics of the intelligent electro-hydraulic hybrid braking system for electric vehicles during a multi-stage collaborative optimization process. Stage 1 reveals the dynamic risk assessment mechanism based on a three-layer weight structure with a fixed weight of 0.3, a linear weight for vehicle speed, and a negative correlation weight for the adhesion coefficient through the time-series curve of the braking risk index. The peak value of 0.79 triggers the regeneration of the braking strategy. Stage 2 uses a bar chart of the electro-hydraulic braking ratio distribution and a comparison chart of the front and rear axle braking force distribution to show the collaborative optimization scheme under high-risk conditions, where the motor braking ratio drops to 35%-40% and the front axle distribution ratio remains stable at 65%-70%. Stage 3 quantifies the differences in ground contact characteristics of each wheel through a bar chart of multi-wheel road adhesion coefficients and a comparison chart of pressure tolerance. The right front wheel adhesion coefficient of 0.54 is 8% higher than that of the left front wheel, and the second pressure tolerance decreases to a margin decay of 1.9 MPa. Stage 4 shows that the gradient descent iteration process curve shows that the algorithm converges to the solution with the lowest braking force fluctuation value within 50ms. Finally, the dynamic response capability of the system under waterlogged road conditions is verified through the braking performance optimization trend line.

[0038] The braking risk index calculation integrates a weighted combination of the first pressure tolerance benchmark value, the second pressure tolerance real-time value, and the braking force deviation. The deviation weight is dynamically adjusted by 0.1 based on the road surface humidity level, reflecting environmental adaptability. During the electro-hydraulic braking ratio optimization process, the gradient descent algorithm divides the three-dimensional braking force distribution space into three subdomains for parallel search: X-axis front axle distribution ratio 60%-80%, Y-axis rear axle distribution ratio 20%-40%, and Z-axis electro-hydraulic ratio 30%-50%. The objective function is to minimize the root mean square difference of the four-wheel braking force. The final output scheme A, with 41% motor braking and 59% hydraulic braking, converges to an acceptable range in terms of fluctuation value. The adhesion coefficient mapping table is pre-calibrated based on the vehicle's full load weight and tire model. When the adhesion coefficient of the right front wheel rises to 0.54, the system automatically adjusts the safety boundary node of that wheel to 3.8 MPa and generates a braking safety domain containing asymmetric distribution schemes using the convex hull algorithm. The pressure tolerance monitoring module adopts a temperature compensation mechanism. In low-temperature environments, it corrects the second pressure tolerance based on the brake fluid viscosity detection value to ensure that the pressure response delay of the hydraulic unit is always controlled within the 100ms threshold.

[0039] Braking performance verification curves show that the system can maintain a braking distance consistency of 92% even under flooded road conditions. The hydraulic adjustment unit, through a PID controller, increases the duty cycle of the inlet valve to 65%, while the electric motor braking unit establishes 285 Nm of regenerative torque within 50 ms, achieving millisecond-level coordination of electro-hydraulic braking force. The entire optimization process generates a braking operation log containing timestamps, electro-hydraulic ratio parameters, and environmental characteristic data, providing data support for subsequent system calibration.

[0040] Example 3: When the braking strategy collaborative generation module calculates the collaborative execution interference degree, it initiates the state switching analysis process. The electro-hydraulic actuator state switching process is divided into a discrete-time step sequence. The command response sequence is captured by the controller area network bus, recording all control commands required to switch from the current real-time electro-hydraulic braking ratio to the reference electro-hydraulic braking ratio, including the torque gradient command of the motor braking unit and the valve opening command sequence of the hydraulic adjustment unit. The pressure adjustment sequence is acquired by a high-frequency pressure sensor in the braking pipeline, recording the pressure transient change curve during the state switching at a sampling rate of 1kHz. The collaborative execution interference degree calculation adopts a two-dimensional evaluation model. The first dimension, command response overlap, is defined as the ratio of the total number of commands N required for state switching to the actual number of reusable commands M. Reusable commands refer to commands that do not need to be resent when the difference between the current actuator state parameters and the reference state parameters is less than a threshold. The second dimension, pressure adjustment amplitude, is taken as the maximum instantaneous value P_max of the second pressure tolerance during the switching process. This value reflects the transient impact intensity of the hydraulic system. Collaborative Execution Interference Degree Calculated using the normalized weighted formula:

[0041] in: and Preset weighting coefficients, default configuration =0.6、 =0.4; Maximum pressure regulation amplitude This is the scaling factor for the system's rated pressure, with a value of 20 MPa. The ratio reflects the efficiency of instruction multiplexing; a higher ratio indicates a smoother state transition process. The relative intensity of quantified pressure fluctuations is considered; a lower ratio indicates better hydraulic system stability. Normalization ensures... The value falls within the interval [0,1]. A value close to 0 indicates low interference, while a value close to 1 indicates high interference. The weighting coefficient is dynamically adjusted based on braking environment characteristics; when the road surface unevenness level exceeds level 3, it is automatically adjusted. Increased to 0.6 to strengthen pressure fluctuation constraints.

[0042] Command response sequence analysis employs command difference matrix technology, constructing an n×m parameter matrix to compare all control parameter differences between the current state and the baseline state. Matrix rows correspond to control parameter types (e.g., motor torque command value, hydraulic valve duty cycle, etc.), and columns represent time series nodes. Two thresholds are set for difference detection: commands with a difference less than 5% are marked as reusable; commands with a difference between 5% and 15% are marked as requiring correction; and commands with a difference exceeding 15% are marked as newly added commands. The actual number of reusable commands, M, is the total number of command nodes with differences less than 5%. The total number of commands N required for state switching includes reusable commands, correction commands, and newly added commands. The analysis process uses a sliding window comparison algorithm with a window width of 100ms to match the control system's response cycle. Pressure regulation sequence processing employs envelope extraction technology. The raw pressure sampling data is first low-pass filtered (cutoff frequency 50Hz) to eliminate high-frequency noise, and then the pressure fluctuation envelope is extracted using Hilbert transform. The maximum value of the second pressure tolerance is also considered. The peak value is taken from the envelope, and the occurrence time of the peak value is recorded simultaneously. To eliminate single-point sampling error, a peak value confirmation rule is set: a valid peak value requires the pressure values ​​of three consecutive sampling points to remain above 98% of the peak value level. The pressure regulation amplitude analysis module is equipped with an outlier rejection mechanism; when a transient fault of the pressure sensor is detected, the previous cycle's value is automatically used. Value substitution. The state transition process is monitored in stages, dividing the transition cycle into three phases: preparation (0-50ms), transition (50-150ms), and stabilization (150-200ms). Command response overlap is calculated independently for each phase. During the preparation phase, the focus is on the reuse rate of motor torque commands; during the transition phase, the reuse rate of hydraulic valve commands is monitored; and during the stabilization phase, the overall command execution completion rate is assessed. Pressure regulation amplitude monitoring focuses on the transition phase, with a 150ms time window specifically set to capture the maximum pressure fluctuation during this period. This staged processing accurately locates interference sources; when the command reuse rate during the preparation phase is detected to be below 70%, command compression optimization is automatically triggered.

[0043] The collaborative interference calculation engine runs within the real-time computing unit, employing a dual-buffer design to ensure data continuity. The main buffer processes real-time data for the current switching cycle, while the auxiliary buffer preloads baseline state parameters for the next cycle. Before calculation begins, data validity is verified. If the instruction sequence missing rate exceeds 10% or the number of invalid stress data sampling points exceeds 5%, calculation is delayed and data retransmission is requested. The calculated result, δ, is output along with a quality flag, categorized into three levels (A, B, and C) based on data integrity and calculation confidence. Only level A results are used by subsequent modules. A rolling storage mechanism is established for historical interference data, retaining the most recent 20 state transitions. Value records are used for trend analysis. When three consecutive values ​​are recorded... A value exceeding 0.8 triggers a system performance alarm, forcibly reducing the baseline state switching frequency by 50%. In low-temperature environments (<-10℃), the interference compensation coefficient is automatically enabled, ultimately... The value is multiplied by a compensation factor of 0.9 to offset the effect of increased brake fluid viscosity. When the calculation results are transmitted to the intervention evaluation value calculation unit, a timestamp and actuator state snapshot are appended to ensure that subsequent evaluation stages can trace the specific operating conditions in which the interference occurred. A safety interruption mechanism is set up during the state switching process; when the pressure regulation amplitude is monitored in real time... The switching process is immediately suspended when the system safety limit is exceeded by 80%. After the interruption, the controller reverts to the previous stable state, replans the switching path, and adds transition steps. A detailed diagnostic report is generated for each interruption event, recording the stress curve characteristics and instruction execution progress at the time of the interruption, which is used to optimize the robustness design of subsequent switching strategies. The system periodically performs offline simulation optimization of the state switching path, trains more efficient instruction sequence combination patterns through machine learning algorithms, and gradually reduces the level of cooperative execution interference under typical operating conditions.

[0044] Example 4: When the braking strategy collaborative generation module calculates the intervention evaluation value of each braking allocation scheme on vehicle stability, a state switching execution flag is introduced. This flag is a three-state variable, with values ​​derived from the enumeration set {immediate execution, delayed execution, no execution}. Immediate execution indicates that the electro-hydraulic actuator state switching must be performed at present; delayed execution indicates that the switching can be performed after 100ms; no execution indicates that the current state is maintained. The intervention evaluation value is calculated using a multi-source data fusion method. The comprehensive braking risk value is taken from the real-time output of the risk assessment module, the collaborative execution interference degree is taken from the latest calculation result of the state switching analysis unit, the path length corresponding to the braking allocation scheme is obtained through vehicle dynamics model simulation, and the road surface smoothness is extracted from the braking environment feature information. When the state switching execution flag is set to immediate execution, the real-time calculated collaborative execution interference degree is directly used for evaluation; when the flag is set to delayed execution, the baseline interference degree is used instead, which is the moving average of the interference degrees of the last 10 normal switching events; when the flag is set to no execution, the collaborative execution interference degree is set to zero. The intervention evaluation value is ultimately expressed as the ratio of the comprehensive braking risk value, the degree of interference in coordinated execution, and the weighted product of the path length and road surface smoothness corresponding to the braking distribution scheme. The lower the ratio, the less intervention the scheme has on vehicle stability.

[0045] The braking control update module modifies the output commands of the hydraulic adjustment unit and the electric motor braking unit based on the electro-hydraulic distribution ratio in the optimized braking control strategy. The hydraulic adjustment unit uses a proportional-integral-derivative controller to adjust the solenoid valve opening, and the target pressure value is obtained by converting the hydraulic braking ratio in the distribution scheme. The electric motor braking unit converts the electric motor braking ratio into a specific regenerative braking torque value through a torque mapping table. Output command transmission adopts a priority scheduling mechanism, with hydraulic commands sent first to ensure basic braking force, and electric motor commands sent within 50ms after the hydraulic command is confirmed. Braking operation log information is recorded in a structured storage format. The strategy generation time is taken from the precise timestamp of the vehicle's timing system. The adopted braking distribution scheme stores its complete parameter set. The braking condition feature set saves data from the most recent 10 cycles in compressed format. Braking environment feature information records the spatiotemporal changes in road surface unevenness and humidity levels. The electro-hydraulic actuator status information includes a comparison of status parameters before and after switching. Refer to Table 1 for the calculation of the intervention evaluation value of the braking distribution scheme generated during a certain braking process.

[0046] Table 1: Calculation of Intervention Evaluation Values ​​for Braking Distribution Scheme

[0047] The state switching execution flag is determined based on the current vehicle stability margin. When the yaw rate deviation exceeds 2.5 degrees / second, it is set to immediate execution; when the deviation is between 1.5 and 2.5 degrees / second, it is set to delayed execution; and when the deviation is below 1.5 degrees / second, it is set to no execution required. During the intervention evaluation value calculation, the weighting coefficients of path length and road surface smoothness are dynamically adjusted according to the vehicle load. Ultimately, scheme A, with the lowest intervention evaluation value, is selected as the optimal electro-hydraulic composite braking distribution scheme. The braking control update module updates the control parameters of scheme A. The target pressure value received by the hydraulic adjustment unit is 8.2 MPa, and the coordinated action of the inlet and outlet valves is adjusted through the PID controller. The torque command value received by the motor braking unit is 320 Nm, and the motor controller uses a ramp-gradient method to achieve torque output. After the output command is sent, the module continuously monitors the execution status. The actual hydraulic pressure reaches 95% of the target value within 150 ms, and the motor torque stabilizes within 98% of the command value within 200 ms. The braking operation log records the generation time of this strategy as 2023-08-15 14:25:36.125. The stored Scheme A parameters include electro-hydraulic ratio, pressure target value, torque command value, and transition time parameters. The braking condition feature set saves key features such as slip ratio 0.18 and adhesion coefficient 0.85. The braking environment feature information records that the road surface unevenness level was level 2 and the humidity level was level 1 during this period. The electro-hydraulic actuator status information includes detailed comparison data of the real-time ratio before switching (motor 30% / hydraulic 70%) and the target ratio after switching (motor 45% / hydraulic 55%). The intervention evaluation value calculation is set with a dynamic verification mechanism, which triggers recalculation when the deviation between the path length and the actual vehicle displacement exceeds 15%. The road surface smoothness data is processed using a moving average filter to eliminate fluctuations caused by single-point acquisition anomalies. The state switching execution flag enables special processing logic under curve conditions, and is forcibly set to immediate execution mode when the steering angle exceeds 90 degrees. The braking control update module sets up a feedback monitoring loop after the command is sent, and checks the deviation between the actual execution value and the target value every 20ms. When the hydraulic pressure deviation exceeds 0.5MPa for 100ms, the pressure compensation program is started. When the motor torque deviation exceeds 15% for 80ms, the torque correction process is triggered.

[0048] Log information storage employs a cyclic overwrite strategy, retaining the most recent 1000 braking operation records, each occupying 256 bytes of storage space. Checksum fields are added to the recorded data to ensure integrity, and differential encoding compression is used during transmission to reduce bandwidth consumption. Under extreme operating conditions, the braking control update module activates a degraded processing mode, automatically switching to the latest effective scheme cached locally when the communication latency exceeds 50ms, ensuring the real-time response capability of the braking system. The system periodically performs offline analysis of historical logs, extracting optimized control parameters under typical operating conditions to update the empirical value settings in the default control strategy.

[0049] Example 5: After the vehicle completes the braking process, the braking performance verification module activates the log parsing program to extract braking operation log information for a specified time period from the on-board storage. The log data packets are sorted according to the strategy generation time. The module first locates the start timestamp of the target braking event, 2023-11-07 09:15:23, and loads the braking distribution scheme parameters, braking condition feature set, and environmental feature information recorded at that time. During the braking process inversion, a digital twin model is constructed. Characteristic parameters such as the road surface adhesion coefficient (0.78) and tire slip ratio (0.22) from the logs are input into the vehicle dynamics simulator. The initial speed is set to 85 km / h as recorded in the logs, and the electro-hydraulic ratio specified in Scheme B (48% motor / 52% hydraulic) is applied. The simulator runs for 200 cycles with a step size of 10 ms, outputting a theoretical braking curve including the speed decay trajectory and a theoretical braking distance of 32.7 meters.

[0050] The actual braking distance is calculated by counting pulses from the wheel speed sensors, tracking the pulse changes from the moment the braking command is issued until the vehicle speed drops to 5 km / h, and then converting this to an actual displacement of 29.8 meters based on the tire rolling radius. The consistency calculation uses the relative error formula: (actual value - theoretical value) / theoretical value × 100%, yielding a deviation of -8.9%. The module automatically associates environmental characteristic information during the braking event; data for road surface unevenness level 2 and humidity level 1 are marked in the evaluation report. The braking performance evaluation report is generated as a PDF document. The first page displays a comparison table of key indicators, the second page includes a speed-time curve comparison graph, and the third page lists detailed braking parameters. When the consistency is detected to be below the preset threshold of -5%, the system stamps a red "Recalibration Required" electronic seal on the report cover. The recalibration command is generated using event coding rules; the command header includes the trigger reason code E005 (insufficient consistency) and the original braking event timestamp. The instruction body specifies the update scope: the feature extraction algorithm version of the braking state perception module needs to be upgraded from V1.2 to V1.3, and the tolerance coefficient of the pressure tolerance calculation parameters of the braking risk assessment module needs to be adjusted. The instructions are transmitted in a time-division manner via the safety bus. The feature extraction algorithm update package, containing the new slip ratio calculation function library, is first sent to the flash memory write area of ​​the perception module; the risk calculation parameter update uses an incremental transmission method, sending only the changed tolerance mapping table fields. After receiving the instructions, both modules enter standby mode and perform the flashing operation the next time the vehicle is stationary.

[0051] The feature extraction algorithm has been upgraded with modifications to the slip ratio calculation process. The original version used a simple ratio of wheel speed to inertial unit speed; the new version adds a steering angle compensation factor: when the steering wheel angle exceeds 30 degrees, the inner wheel speed is dynamically corrected according to the steering geometry model. The road surface adhesion coefficient extraction algorithm has added multi-sensor fusion weights, increasing the weight of millimeter-wave radar data from 0.6 to 0.7, while the weight of camera data is correspondingly reduced. Risk calculation parameters have been updated to focus on the load sensitivity of the first pressure tolerance; the original full-load correction coefficient of 1.2 has been adjusted to 1.15, and a low-temperature compensation item has been added: when the brake fluid temperature is below -5℃, the tolerance value automatically increases by 0.3MPa. During the calibration and verification phase, a test case set was injected, including six typical working conditions such as dry asphalt pavement (adhesion coefficient 0.85) and wet cement pavement (adhesion coefficient 0.45). Each test case underwent twenty simulated braking operations, and consistency data was collected to establish a statistical distribution chart. When the mean consistency of all test cases falls within the range of -3% to +2%, the calibration result is considered valid. The first actual braking after the new parameters took effect occurred at 14:30:17 on November 8, 2023. The system recorded an actual braking distance of 31.2 meters, compared to the theoretical value of 31.5 meters, improving the consistency to -0.95%. The performance verification module is set to perform periodic self-checks, automatically initiating the verification process every 50 braking events or every 72 hours. Immediate verification is forcibly triggered after special conditions such as long downhill sections, regardless of whether the consistency exceeds the standard. A historical evaluation report index database is established, supporting multi-dimensional searches by date range, road surface type, or braking intensity. When the same road segment experiences three consecutive instances of abnormal consistency, the system automatically labels the characteristic combination of that segment and generates a special observation list.

[0052] The recalibration process incorporates a safety protection mechanism. Before parameter flashing, a backup image of the original configuration is created, and three sets of comparative tests are executed after flashing. If any test case experiences system instability or control failure, it immediately rolls back to the previous stable version. Calibration command transmission employs a dual-verification protocol, with the receiving end returning parameter hash values ​​for the sending end to compare and confirm. During calibration execution, the braking system switches to pure hydraulic backup mode, and the electric motor braking unit is disabled to ensure safety redundancy. After all updates are completed, a calibration report is generated, detailing the changes, test results, and effective dates. This report, along with the braking performance evaluation report, is archived in the vehicle lifecycle database.

[0053] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An intelligent electro-hydraulic hybrid braking system for electric vehicles, characterized in that, The system includes: The braking state perception module collects vehicle operating status data and road environment information in real time through multi-source sensors to construct a braking condition feature set and braking environment feature information. The braking risk assessment module, based on braking environment characteristic information, combined with the vehicle's initial braking parameters, current braking request, and current time electro-hydraulic actuator status information, obtains a comprehensive braking risk value based on a braking condition feature set. The braking strategy collaborative generation module generates multiple electro-hydraulic composite braking allocation schemes based on the braking condition feature set when the comprehensive braking risk value is abnormal. Combining the braking environment feature information, it calculates the collaborative execution interference degree corresponding to the current time when the electro-hydraulic actuator state switches to the reference state, obtains the intervention evaluation value of each braking allocation scheme on vehicle stability, selects the best electro-hydraulic composite braking allocation scheme, and generates the optimized braking control strategy for the current time. The braking control update module updates the braking control parameters in real time according to the optimized braking control strategy at the current time and generates braking operation log information.

2. The intelligent electro-hydraulic hybrid braking system for electric vehicles according to claim 1, characterized in that, The multi-source sensors include wheel speed sensors, pressure sensors, inertial measurement units, and road surface recognition sensors; Each element in the braking condition feature set corresponds to a braking condition feature information, which includes tire slip ratio, braking force distribution coefficient and road surface adhesion coefficient extracted based on sensor data. The braking environment characteristics information includes road surface unevenness and humidity levels in different areas.

3. The intelligent electro-hydraulic hybrid braking system for electric vehicles according to claim 2, characterized in that, The initial braking parameters of the vehicle include a reference electro-hydraulic braking ratio and a first pressure tolerance. The current time electro-hydraulic actuator status information includes the real-time electro-hydraulic braking ratio and the second pressure tolerance; The first pressure tolerance represents the maximum value of the difference between the rated pressure and the actual pressure of the brake line under reference conditions; The second pressure tolerance represents the maximum value of the difference between the rated pressure and the actual pressure of the brake line under the current condition.

4. The intelligent electro-hydraulic hybrid braking system for electric vehicles according to claim 3, characterized in that, When the braking risk assessment module calculates the comprehensive braking risk value, it records the comprehensive braking risk value of the vehicle based on the braking condition feature set as the braking risk index. The braking risk index is obtained by a weighted combination of the first pressure tolerance, the second pressure tolerance, and the deviation of ideal braking force from the current braking request and braking condition characteristics. When the actual braking force is higher than the ideal braking force, the deviation is taken as a positive value; otherwise, it is taken as a negative value.

5. The intelligent electro-hydraulic hybrid braking system for electric vehicles according to claim 2, characterized in that, When the braking strategy collaborative generation module generates multiple electro-hydraulic composite braking distribution schemes, it extracts the safe braking force boundary nodes corresponding to each adhesion coefficient in the braking condition feature set, marks the safe boundaries corresponding to each adhesion coefficient in the braking force distribution interval, and records the feasible domain between the current braking force request value and each safe boundary node as the braking distribution safe domain. The brake allocation safety domain includes one or more brake allocation subdomains; The gradient descent algorithm is used to solve the electro-hydraulic hybrid braking allocation scheme corresponding to each braking allocation subdomain.

6. The intelligent electro-hydraulic hybrid braking system for electric vehicles according to claim 5, characterized in that, When the braking strategy collaborative generation module calculates the collaborative execution interference degree, it obtains the command response sequence and pressure regulation sequence during the state switching process of the electro-hydraulic actuator. The degree of interference in coordinated execution is obtained by normalized weighting of instruction response overlap and pressure regulation amplitude. The instruction response overlap is calculated based on the ratio of the total number of instructions required for state switching to the actual number of reusable instructions. The pressure regulation amplitude is taken as the maximum value of the second pressure tolerance during the state switching process.

7. The intelligent electro-hydraulic hybrid braking system for electric vehicles according to claim 6, characterized in that, When the braking strategy collaborative generation module calculates the intervention evaluation value of each braking distribution scheme on vehicle stability, a state switching execution flag is introduced. The intervention evaluation value is obtained by the ratio of the comprehensive braking risk value, the degree of interference in coordinated execution, and the weighted product of the path length and road surface smoothness corresponding to the braking distribution scheme. When the state transition execution flag is true, the cooperative execution interference degree is included in the calculation; otherwise, the baseline interference degree is used instead.

8. The intelligent electro-hydraulic hybrid braking system for electric vehicles according to claim 1, characterized in that, When the braking control update module updates the braking control parameters, it corrects the output commands of the hydraulic adjustment unit and the motor braking unit according to the electro-hydraulic distribution ratio in the optimized braking control strategy. The braking operation log information includes the strategy generation time, the braking allocation scheme adopted, the braking condition feature set, the braking environment feature information, and the electro-hydraulic actuator status information.

9. The intelligent electro-hydraulic hybrid braking system for electric vehicles according to claim 1, characterized in that, The system also includes a braking performance verification module, which inverts the braking process based on braking operation log information, calculates the degree of agreement between the actual braking distance and the theoretical braking distance, and generates a braking performance evaluation report. When the consistency is lower than the threshold, the braking performance verification module triggers a system recalibration command. The recalibration command is sent to the braking state perception module and the braking risk assessment module to update the feature extraction algorithm and risk calculation parameters.

10. A method for intelligent electro-hydraulic hybrid braking control of an electric vehicle, applied to an intelligent electro-hydraulic hybrid braking system for an electric vehicle as described in any one of claims 1 to 9, characterized in that, The method includes the following steps: By collecting vehicle operating status data and road environment information in real time through multi-source sensors, a braking condition feature set and braking environment feature information are constructed. Based on the braking environment feature information, combined with the vehicle's initial braking parameters, current braking request, and current time electro-hydraulic actuator status information, a comprehensive braking risk value based on the braking condition feature set is calculated. When the comprehensive braking risk value is abnormal, multiple electro-hydraulic composite braking distribution schemes are generated based on the braking condition feature set. The cooperative execution interference degree corresponding to the switching of the electro-hydraulic actuator state to the baseline state at the current time is calculated in combination with the braking environment feature information. Then, the intervention evaluation value of each braking distribution scheme on vehicle stability is obtained, and the best electro-hydraulic composite braking distribution scheme is selected to generate the optimized braking control strategy at the current time. The braking control parameters are updated in real time according to the optimized braking control strategy, and braking operation log information is generated.

Images