An EMB Torque Distribution Control Method Based on Road Surface Adhesion Coefficient Estimation Algorithm

Abstract

This invention specifically relates to an EMB torque distribution control method based on a road surface adhesion coefficient estimation algorithm, belonging to the field of vehicle braking control technology. It includes: rolling time-domain adhesion coefficient estimation based on multi-source data fusion; and optimized braking force distribution based on yaw moment balance. In this invention, the rolling time-domain adhesion coefficient estimation method utilizes historical data from cameras, IMUs, GNSS, and wheel speed sensors to construct a nonlinear adaptive observer, overcoming the limitations of traditional single-sensor or current-moment data estimation. On one hand, the camera uses UNet semantic segmentation and a ShuffleNet network to identify road surface types, pre-defining the adhesion coefficient range and reducing initial estimation errors. On the other hand, the rolling time-domain observer combines the vehicle dynamics model with historical data, achieving exponential stable convergence only when the tires enter the nonlinear region at a historical moment, demonstrating strong robustness to model uncertainties.

Description

Technical Field

[0001] This invention relates to the field of vehicle braking control technology, and in particular to an EMB torque distribution control method based on a road surface adhesion coefficient estimation algorithm. Background Technology

[0002] EMB (Electro-mechanical Braking System), as an innovative braking technology, is gradually becoming an important part of modern vehicles. Combining the advantages of electric braking and mechanical control, EMB offers faster response, higher braking accuracy, and a smaller size, making it an ideal replacement for traditional hydraulic braking systems and improving vehicle safety and handling. However, current EMB technology does not adequately consider changes in the road surface adhesion coefficient and differences in the road surface adhesion coefficient between wheels in its torque distribution, which significantly limits its braking capability.

[0003] Commonly used EMB (Electrical Motion Shield) systems often assume a high-friction surface when distributing braking torque. This can easily lead to wheel lock-up and skidding when the vehicle is traveling on a low-friction surface. Some studies consider incorporating road surface adhesion coefficient (OPC) estimation based on vehicle / tire models to improve EMB's adaptability to different road conditions. However, existing methods often only utilize current sensor data to evaluate OPC. Due to model uncertainties and insufficient measurement records, the estimation results tend to be inaccurate and slow in convergence, resulting in poor performance of EMB in practical applications due to OPC mismatch. Other studies consider using camera information to estimate road conditions, but simply classifying the road surface does not yield an accurate OPC. For example, on wet and slippery surfaces, the OPC can vary between 0.3 and 0.9. Therefore, vehicle / tire model-based methods are still needed to further estimate the accurate OPC to improve EMB performance.

[0004] Therefore, an EMB torque distribution control method based on a road adhesion coefficient estimation algorithm is needed. This method utilizes current and past information from cameras, IMUs (Inertial Measurement Units), and GNSS (Global Navigation Satellite System), combined with vehicle dynamics models and road surface classification from cameras. By designing a constrained rolling time-domain nonlinear adaptive observer, the problem of slow estimation speed and low accuracy caused by model uncertainty and measurement excitation can be reduced. This method is then applied to the torque distribution of the EMB system, thereby improving the optimal braking capability of the EMB system. Summary of the Invention

[0005] The purpose of this invention is to propose an EMB torque distribution control method based on a road surface adhesion coefficient estimation algorithm to solve the above-mentioned problems.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] An EMB torque distribution control method based on a road surface adhesion coefficient estimation algorithm includes:

[0008] Rolling temporal adhesion coefficient estimation by multi-source data fusion: A nonlinear adaptive observer is constructed using historical data from cameras, IMUs, GNSS, and wheel speed sensors to output the estimated real-time road adhesion coefficients for each wheel and obtain the optimal slip ratio for each wheel.

[0009] Braking force optimization based on yaw moment balance: The driver's operation is analyzed into the required braking torque and yaw moment; according to the road adhesion coefficient and the optimal slip ratio, combined with the driver's needs and the activation status of ABS, the braking force is distributed, and then the braking torque of each wheel is dynamically compensated, and the final output braking torque is obtained.

[0010] Preferably, the rolling temporal adhesion coefficient estimation of the multi-source data fusion specifically includes road surface adhesion estimation and observer design and performance analysis.

[0011] Preferably, the road surface adhesion estimation specifically includes:

[0012] When the vehicle is driving normally, the camera classifies the road surface by displaying images of the road conditions.

[0013] The semantic segmentation network UNet is used to extract the drivable area of ​​the road surface. Taking into account the amount of information contained in the road image block and the utilization rate of the drivable area, an optimization algorithm is designed to divide the road image block into appropriate blocks.

[0014] A lightweight convolutional neural network, ShuffleNet, is used to extract road surface image features and obtain image-based road surface type recognition results. After obtaining road surface classification information, the range of road surface adhesion coefficient is defined based on the road surface type as […]. .

[0015] Preferably, the observer design and performance analysis specifically include:

[0016] Based on the road surface adhesion range, combined with the vehicle dynamics model and historical data from sensors, a nonlinear adaptive observer for the road surface adhesion coefficient based on the rolling time domain is designed.

[0017] Assuming the current time The wheel speed is The wheel-end braking torque is speed The range of road surface adhesion coefficient is [ ;past , , ..., The wheel speed at time is , , ..., Correspondingly, the wheel-end driving braking torque is , , ..., Vehicle speed information is , , ..., Design an adaptive observer for road surface adhesion coefficient based on historical and current information;

[0018]

[0019] in,

[0020]

[0021]

[0022] , , , and For positive integers, ; … Each represents the current time. and the past ,… time The estimated value; ,… For the past Wheel slip ratio at any given moment; Estimate the road surface adhesion coefficient at the current moment.

[0023] Preferably, the braking force optimization distribution based on yaw moment balance specifically includes EMB torque distribution and EMB system braking force distribution.

[0024] Preferably, the EMB torque distribution specifically includes:

[0025] Based on the total required braking torque and yaw moment, and considering the road adhesion coefficient constraints of each wheel, the braking torque of each wheel is allocated at the cost of minimizing the road adhesion coefficient of all wheels. The specific optimization problem is as follows:

[0026]

[0027]

[0028] Subjectto

[0029] If the ABS activation for each wheel

[0030] In the formula:

[0031] This indicates the braking torque of each wheel;

[0032] The vertical load on each wheel;

[0033] This represents the road adhesion coefficient of each wheel;

[0034] r represents the wheel radius;

[0035] This indicates half the distance between the left and right wheels on both sides of the front and rear axles;

[0036] To meet the yaw moment requirement;

[0037] This indicates that when the ABS of one or more wheels is activated, the corresponding... The braking torque calculated by ABS is not used as an optimization variable in the optimization calculation.

[0038] Preferably, the braking force distribution of the EMB system specifically includes:

[0039] Based on the transmission chain dynamics model, combined with the load torque observer to identify the transmission efficiency attenuation coefficient and nonlinear friction characteristics online, the inertial hysteresis effect of the motor and ball screw is eliminated through the inertial compensation algorithm.

[0040] A combined feedforward and feedback control architecture is adopted. The feedforward link dynamically corrects the target torque based on the road surface adhesion coefficient and wheel speed fluctuation, while the feedback link estimates the disturbance torque in real time through a sliding mode observer and injects compensation current.

[0041] Simultaneously, an adaptive weighting factor is introduced, and data from the wheel cylinder pressure sensor and the motor encoder are integrated to iteratively optimize the compensation amount.

[0042] Preferably, the elimination of the inertial hysteresis effect between the motor and the ball screw through the inertial compensation algorithm specifically includes:

[0043] In the EMB system drivetrain, the motor rotor, ball screw, and brake caliper components have rotational inertia. When the control torque When the system changes, it produces angular acceleration due to inertia. This causes the actual torque output to lag behind the control signal;

[0044] According to the rotational law formula: ;

[0045] in This is the frictional resistance torque;

[0046] Moment of inertia The determination process includes:

[0047] Obtain the mass of the motor rotor ,radius ,length Substitute the formula for moment of inertia: ;

[0048] Obtain the rotational moment of inertia ;

[0049] The vehicle is subjected to emergency braking at preset time intervals to obtain the motor temperature change within a preset duration, and the highest and lowest temperature values ​​are extracted from it.

[0050] The temperature rise is obtained by calculating the difference between the highest and lowest temperature values.

[0051] After obtaining the time points corresponding to the highest and lowest temperature values, extract the duration between the highest and lowest temperature values, divide the temperature rise by the duration between the highest and lowest temperature values ​​to obtain the unit temperature rise value.

[0052] A preset unit temperature rise threshold is used. The unit temperature rise value is divided by the unit temperature rise threshold to obtain the unit temperature rise ratio.

[0053] Preferably, the method further includes:

[0054] The surface of the motor rotor is divided into several sub-regions according to a preset size, and the temperature of each sub-region and the center of the motor rotor is obtained after the vehicle performs emergency braking.

[0055] The surface temperature difference is obtained by subtracting the corresponding center temperature of the motor rotor from the temperature of each sub-region of the motor rotor.

[0056] Extract the maximum surface temperature difference value in the sub-region of the motor rotor and record it as the maximum surface temperature difference value; sequentially obtain the maximum surface temperature difference value of each sub-region of the electronic rotor, sort the maximum surface temperature difference values ​​of each sub-region in descending order of numerical value, and extract the two largest maximum surface temperature differences and their corresponding positions; mark the positions of the two largest maximum surface temperature differences as the first positioning point and the second positioning point, respectively.

[0057] Mark the center position of the rotor as the third positioning point. Connect the first, second, and third positioning points with straight lines to obtain a triangle. Calculate the area of ​​the triangle and record it as the approximation distribution value.

[0058] After normalizing the unit temperature rise ratio and the fitted distribution value, the unit temperature rise ratio and the fitted distribution value are used as the major and minor semi-axes of the ellipse, respectively, to construct an elliptical model. The area of ​​the elliptical model is calculated and recorded as the added inertia value.

[0059] The inertia matching value is obtained by weighting the rotational defined inertia and the added inertia value.

[0060] Match the corresponding moment of inertia based on the inertia matching value, and substitute the obtained moment of inertia into the rotational law formula mentioned above: That's all.

[0061] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0062] 1. This invention employs a rolling temporal-domain adhesion coefficient estimation method based on multi-source data fusion. It utilizes historical data from cameras, IMUs, GNSS, and wheel speed sensors to construct a nonlinear adaptive observer, overcoming the limitations of traditional single-sensor or current-moment data estimation. On one hand, the camera identifies road surface types through UNet semantic segmentation and ShuffleNet networks, pre-defining the adhesion coefficient range and reducing initial estimation errors. On the other hand, the rolling temporal-domain observer combines vehicle dynamics models with historical data, achieving exponential stable convergence only when the tires have entered the nonlinear region at historical moments, demonstrating strong robustness to model uncertainties.

[0063] 2. This invention employs a braking force optimization distribution strategy based on yaw moment balance. It analyzes driver operation into required braking torque and yaw moment, combining the adhesion coefficient constraints of each wheel and the ABS activation state. A convex optimization problem is constructed with the goal of minimizing the road surface adhesion coefficient. Dynamic torque distribution is achieved through OSQP solution, resolving the wheel lock-up problem on low-adhesion surfaces caused by traditional fixed-ratio distribution. Simultaneously, the EMB system's braking force distribution uses a transmission chain dynamics model and a load torque observer to identify friction characteristics online. Combined with the motor rotor temperature distribution characteristics, it dynamically corrects the moment of inertia, improving the accuracy of the inertia compensation torque. Attached Figure Description

[0064] Further details, features, and advantages of this application are disclosed in the following description of exemplary embodiments in conjunction with the accompanying drawings, in which:

[0065] Figure 1 This is a flowchart of the present invention. Detailed Implementation

[0066] Several embodiments of this application will now be described in more detail with reference to the accompanying drawings to enable those skilled in the art to implement this application. This application may be embodied in many different forms and for various purposes and should not be limited to the embodiments set forth herein. These embodiments are provided to make this application thorough and complete, and to fully convey the scope of this application to those skilled in the art. The embodiments described do not limit this application.

[0067] Unless otherwise defined, all terms used herein (including technical and scientific terms) shall have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. It will be further understood that terms such as those defined in commonly used dictionaries shall be interpreted as having a meaning consistent with their meaning in the relevant field and / or the context of this specification, and shall not be interpreted in an idealized or overly formal sense unless expressly defined herein.

[0068] Please see Figure 1 As shown, the present invention provides a technical solution:

[0069] An EMB torque distribution control method based on a road surface adhesion coefficient estimation algorithm includes:

[0070] Rolling temporal adhesion coefficient estimation by multi-source data fusion: A nonlinear adaptive observer is constructed using historical data from cameras, IMUs, GNSS, and wheel speed sensors to output the estimated real-time road adhesion coefficients for each wheel and obtain the optimal slip ratio for each wheel.

[0071] Specifically, this includes road surface adhesion estimation, as well as observer design and performance analysis;

[0072] Road surface adhesion estimation, specifically including:

[0073] When the vehicle is driving normally, the camera classifies the road surface by displaying images of the road conditions.

[0074] The semantic segmentation network UNet is used to extract the drivable area of ​​the road surface. Taking into account the amount of information contained in the road image block and the utilization rate of the drivable area, an optimization algorithm is designed to divide the road image block into appropriate blocks.

[0075] A lightweight convolutional neural network, ShuffleNet, is used to extract road surface image features and obtain image-based road surface type recognition results. After obtaining road surface classification information, the range of road surface adhesion coefficient is defined based on the road surface type as […]. ;

[0076] Observer design and performance analysis, specifically including:

[0077] Based on the road surface adhesion range, combined with the vehicle dynamics model and historical data from sensors, a nonlinear adaptive observer for the road surface adhesion coefficient based on the rolling time domain is designed.

[0078] Assuming the current time The wheel speed is The wheel-end braking torque is speed The range of road surface adhesion coefficient is [ ;past , , ..., The wheel speed at time is , , ..., Correspondingly, the wheel-end driving braking torque is , , ..., Vehicle speed information is , , ..., Design an adaptive observer for road surface adhesion coefficient based on historical and current information;

[0079]

[0080] in,

[0081]

[0082]

[0083] , , , and For positive integers, ; … Each represents the current time. and the past ,… time The estimated value; ,… For the past Wheel slip ratio at a given moment. Estimate the road surface adhesion coefficient at the current moment;

[0084] It is worth noting that in this observer, for each moment, it is only necessary to ensure that the past... At any given moment, if the tire enters the nonlinear region (i.e., it does not need to remain in the nonlinear region), the observer can converge exponentially, thus ensuring the robustness of the estimation results to model uncertainties.

[0085] Furthermore, if the tires are always operating in the nonlinear region, the convergence speed of the observer is proportional to the time domain length of the collected data, which also reflects that the designed observer can accelerate the convergence of the estimation. At the same time, since the camera is used to limit the initial value of the road adhesion estimation and its subsequent corresponding estimation range, the road adhesion estimation results can also be accelerated to a certain extent. Finally, since the observer can be applied to each wheel, the road adhesion coefficient of each wheel at each time can be obtained.

[0086] Braking force optimization based on yaw moment balance: The driver's operation (such as brake pedal force and steering input) is analyzed into the required braking torque and yaw moment; according to the road adhesion coefficient and the optimal slip ratio, combined with the driver's needs and the activation status of ABS, the braking force is distributed, and then the braking torque of each wheel is dynamically compensated, and the final output braking torque is obtained.

[0087] Specifically, this includes EMB torque distribution and EMB system braking force distribution;

[0088] EMB torque distribution specifically includes:

[0089] Based on the total required braking torque and yaw moment, and considering the road adhesion coefficient constraints of each wheel, the braking torque of each wheel is allocated at the cost of minimizing the road adhesion coefficient of all wheels. The specific optimization problem is as follows:

[0090]

[0091]

[0092] Subjectto

[0093] If the ABS activation for each wheel

[0094] In the formula:

[0095] This indicates the braking torque of each wheel;

[0096] The vertical load on each wheel;

[0097] This represents the road adhesion coefficient of each wheel;

[0098] r represents the wheel radius;

[0099] This indicates half the distance between the left and right wheels on both sides of the front and rear axles;

[0100] To meet the yaw moment requirement;

[0101] This indicates that when the ABS of one or more wheels is activated, the corresponding... The braking torque calculated by ABS is not used as an optimization variable in the optimization calculation.

[0102] The cost function and constraints described above constitute the optimization problem of braking torque distribution. Since this optimization problem is a convex optimization problem, the braking torque of each wheel is solved using OSQP.

[0103] The following further explains the constraints on ABS-activated wheels. When allocating braking torque optimally, the activation of the ABS system must be considered. If the ABS at the wheel end of one or more wheels is activated, the braking torque of that wheel is governed by its ABS and does not participate in the optimal allocation of braking torque by the EMB system.

[0104] For example, if the ABS of the first wheel is activated, then the braking torque of the first wheel is equal to the braking torque provided by the ABS of the first wheel. The optimal distribution of braking torque of the EMB system can only be distributed to the remaining three wheels.

[0105] The EMB system's braking force distribution specifically includes:

[0106] Based on the transmission chain dynamics model, combined with the load torque observer to identify the transmission efficiency attenuation coefficient and nonlinear friction characteristics online, the inertial hysteresis effect of the motor and ball screw is eliminated through the inertial compensation algorithm.

[0107] A combined feedforward and feedback control architecture is adopted. The feedforward link dynamically corrects the target torque based on the road surface adhesion coefficient and wheel speed fluctuation, while the feedback link estimates the disturbance torque in real time through a sliding mode observer and injects compensation current.

[0108] Simultaneously, an adaptive weighting factor is introduced, and data from the wheel cylinder pressure sensor and the motor encoder are fused to iteratively optimize the compensation amount.

[0109] The inertial hysteresis effect between the motor and the ball screw is eliminated through an inertial compensation algorithm, specifically including:

[0110] In the EMB system drivetrain, the motor rotor, ball screw, and brake caliper components have rotational inertia. (unit: When the control torque When the system changes, it produces angular acceleration due to inertia. This causes the actual torque output to lag behind the control signal;

[0111] According to the rotational law formula: ;

[0112] in This is the frictional resistance torque;

[0113] Moment of inertia The determination process includes:

[0114] Obtain the mass of the motor rotor ,radius ,length Substitute the formula for moment of inertia: ;

[0115] Obtain the rotational moment of inertia ;

[0116] The vehicle is subjected to emergency braking at preset time intervals to obtain the motor temperature change within a preset duration, and the highest and lowest temperature values ​​are extracted from the results; the lowest temperature value is a fixed value, and the highest temperature value is the highest temperature obtained from multiple experiments.

[0117] The temperature rise is obtained by calculating the difference between the highest and lowest temperature values.

[0118] After obtaining the time points corresponding to the highest and lowest temperature values, extract the duration between the highest and lowest temperature values, divide the temperature rise by the duration between the highest and lowest temperature values ​​to obtain the unit temperature rise value.

[0119] A preset unit temperature rise threshold is used to divide the unit temperature rise value by the unit temperature rise threshold to obtain the unit temperature rise ratio.

[0120] The surface of the motor rotor is divided into several sub-regions according to a preset size, and the temperature of each sub-region and the center of the motor rotor is obtained after the vehicle performs emergency braking.

[0121] The surface temperature difference is obtained by subtracting the corresponding center temperature of the motor rotor from the temperature of each sub-region of the motor rotor.

[0122] Extract the maximum surface temperature difference value in the sub-region of the motor rotor and record it as the maximum surface temperature difference value; sequentially obtain the maximum surface temperature difference value of each sub-region of the electronic rotor, sort the maximum surface temperature difference values ​​of each sub-region in descending order of numerical value, and extract the two largest maximum surface temperature differences and their corresponding positions; mark the positions of the two largest maximum surface temperature differences as the first positioning point and the second positioning point, respectively.

[0123] Mark the center position of the rotor as the third positioning point. Connect the first, second, and third positioning points with straight lines to obtain a triangle. Calculate the area of ​​the triangle and record it as the approximation distribution value.

[0124] After normalizing the unit temperature rise ratio and the fitted distribution value, the unit temperature rise ratio and the fitted distribution value are used as the major and minor semi-axes of the ellipse, respectively, to construct an elliptical model. The area of ​​the elliptical model is calculated and recorded as the added inertia value.

[0125] The inertia matching value is obtained by weighting the rotational defined inertia and the added inertia value.

[0126] The weighting factors for the predefined rotational inertia and the added inertia value are calculated by multiplying the defined rotational inertia and the added inertia value with their corresponding weighting factors, and then summing them to obtain the inertia matching value.

[0127] Match the corresponding moment of inertia based on the inertia matching value, and substitute the obtained moment of inertia into the rotational law formula mentioned above: That's all;

[0128] Moment of inertia: Preset multiple threshold ranges, and each threshold range corresponds to a moment of inertia. Match the moment of inertia matching value with the ranges of multiple thresholds to obtain the moment of inertia corresponding to the moment of inertia matching value.

[0129] The rotor surface of the motor is divided into sub-regions, and the moment of inertia is corrected by combining temperature distribution characteristics:

[0130] On the one hand, it breaks through the limitations of traditional average temperature calculation, captures the spatial non-uniformity of rotor thermal deformation through sub-region temperature difference analysis, and uses triangular approximation distribution values ​​to quantify the influence of temperature gradient on inertia distribution.

[0131] On the other hand, by combining the unit temperature rise ratio to dynamically reflect the rate of temperature change, the thermophysical properties and geometric features are coupled through an elliptical model, so that the added inertia value can accurately characterize the inertia fluctuation under thermo-mechanical coupling. Finally, the dynamic calibration of the rotational inertia is achieved through weighted calculation and threshold matching, which not only improves the adaptability of the inertia compensation algorithm to temperature field changes, but also ensures real-time calculation efficiency, thereby enhancing the torque response accuracy and stability of the EMB system under complex braking conditions.

[0132] The above formulas are derived from software simulations using a large amount of data and are selected to be close to the actual values. The influence weight factors and specific coefficient values ​​in the formulas are set by those skilled in the art based on the actual situation and can be adjusted and modified in the future.

[0133] The above description of the embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An EMB torque distribution control method based on a road surface adhesion coefficient estimation algorithm, characterized in that, include: Rolling temporal adhesion coefficient estimation by multi-source data fusion: A nonlinear adaptive observer is constructed using historical data from cameras, IMUs, GNSS, and wheel speed sensors to output the estimated real-time road adhesion coefficients for each wheel and obtain the optimal slip ratio for each wheel. Braking force optimization based on yaw moment balance: The driver's operation is analyzed into the required braking torque and yaw moment; Based on the road surface adhesion coefficient and optimal slip ratio, combined with the driver's needs and the activation status of ABS, braking force is distributed, and then the braking torque of each wheel is dynamically compensated to finally output the braking torque. The EMB system's braking force distribution specifically includes: Based on the transmission chain dynamics model, combined with the load torque observer to identify the transmission efficiency attenuation coefficient and nonlinear friction characteristics online, the inertial hysteresis effect of the motor and ball screw is eliminated through the inertial compensation algorithm. A combined feedforward and feedback control architecture is adopted. The feedforward link dynamically corrects the target torque based on the road surface adhesion coefficient and wheel speed fluctuation, while the feedback link estimates the disturbance torque in real time through a sliding mode observer and injects compensation current. Simultaneously, an adaptive weighting factor is introduced, and data from the wheel cylinder pressure sensor and the motor encoder are fused to iteratively optimize the compensation amount. In the EMB system drive train, the motor rotor, ball screw, and brake caliper components have rotational inertia. When the control torque changes, the system generates angular acceleration due to inertia, causing the actual torque output to lag behind the control signal. The rotational moment of inertia is thus calculated accordingly. The vehicle is subjected to emergency braking at preset time intervals to obtain the motor temperature change within a preset duration, and the highest and lowest temperature values ​​are extracted from it. The temperature rise is obtained by calculating the difference between the highest and lowest temperature values. After obtaining the time points corresponding to the highest and lowest temperature values, extract the duration between the highest and lowest temperature values, divide the temperature rise by the duration between the highest and lowest temperature values ​​to obtain the unit temperature rise value. A preset unit temperature rise threshold is used. The unit temperature rise value is divided by the unit temperature rise threshold to obtain the unit temperature rise ratio.

2. The EMB torque distribution control method based on a road surface adhesion coefficient estimation algorithm according to claim 1, characterized in that, The rolling temporal adhesion coefficient estimation of the multi-source data fusion specifically includes road surface adhesion estimation as well as observer design and performance analysis.

3. The EMB torque distribution control method based on a road surface adhesion coefficient estimation algorithm according to claim 2, characterized in that, Road surface adhesion estimation, specifically including: When the vehicle is driving normally, the camera classifies the road surface by displaying images of the road conditions. The semantic segmentation network UNet is used to extract the drivable area of ​​the road surface. Taking into account the amount of information contained in the road image block and the utilization rate of the drivable area, an optimization algorithm is designed to divide the road image block into appropriate blocks. A lightweight convolutional neural network, ShuffleNet, is used to extract road surface image features and obtain image-based road surface type recognition results. After obtaining road surface classification information, the range of road surface adhesion coefficient is defined based on the road surface type as […]. .

4. The EMB torque distribution control method based on a road surface adhesion coefficient estimation algorithm according to claim 3, characterized in that, Observer design and performance analysis, specifically including: Based on the road surface adhesion range, combined with the vehicle dynamics model and historical data from sensors, a nonlinear adaptive observer for the road surface adhesion coefficient based on the rolling time domain is designed. Assuming the current time The wheel speed is The wheel-end braking torque is speed The range of road surface adhesion coefficient is [ ;past , , ..., The wheel speed at time is , , ..., Correspondingly, the wheel-end driving braking torque is , , ..., Vehicle speed information is , , ..., Design an adaptive observer for road surface adhesion coefficient based on historical and current information; ， in, ， ， , , , and For positive integers, ; … Each represents the current time. and the past ,… time The estimated value; ,… For the past Wheel slip ratio at any given moment; Estimate the road surface adhesion coefficient at the current moment.

5. The EMB torque distribution control method based on a road surface adhesion coefficient estimation algorithm according to claim 1, characterized in that, Braking force optimization based on yaw moment balance includes EMB torque distribution and EMB system braking force distribution.

6. The EMB torque distribution control method based on a road surface adhesion coefficient estimation algorithm according to claim 5, characterized in that, EMB torque distribution specifically includes: Based on the total required braking torque and yaw moment, and considering the road adhesion coefficient constraints of each wheel, the braking torque of each wheel is allocated at the cost of minimizing the road adhesion coefficient of all wheels. The specific optimization problem is as follows: ， ， Subjectto , If the Activate the ABS on each wheel. In the formula: This indicates the braking torque of each wheel; The vertical load on each wheel; This represents the road adhesion coefficient of each wheel; r represents the wheel radius; This indicates half the distance between the left and right wheels on both sides of the front and rear axles; To meet the yaw moment requirement; This indicates that when the ABS of one or more wheels is activated, the corresponding... The braking torque calculated by ABS is not used as an optimization variable in the optimization calculation.

7. The EMB torque distribution control method based on a road surface adhesion coefficient estimation algorithm according to claim 1, characterized in that, Also includes: The surface of the motor rotor is divided into several sub-regions according to a preset size, and the temperature of each sub-region and the center of the motor rotor is obtained after the vehicle performs emergency braking. The surface temperature difference is obtained by subtracting the corresponding center temperature of the motor rotor from the temperature of each sub-region of the motor rotor. Extract the maximum surface temperature difference value in the sub-region of the motor rotor and record it as the maximum surface temperature difference value; sequentially obtain the maximum surface temperature difference value of each sub-region of the electronic rotor, sort the maximum surface temperature difference values ​​of each sub-region in descending order of numerical value, and extract the two largest maximum surface temperature differences values ​​and their corresponding positions. Mark the locations of the two largest surface temperature differences as the first positioning point and the second positioning point, respectively. Mark the center position of the rotor as the third positioning point. Connect the first, second, and third positioning points with straight lines to obtain a triangle. Calculate the area of ​​the triangle and record it as the approximation distribution value. After normalizing the unit temperature rise ratio and the fitted distribution value, the unit temperature rise ratio and the fitted distribution value are used as the major and minor semi-axes of the ellipse, respectively, to construct an elliptical model. The area of ​​the elliptical model is calculated and recorded as the added inertia value. The inertia matching value is obtained by weighting the rotational defined inertia and the added inertia value. Match the corresponding moment of inertia to the inertia matching value, and substitute the obtained moment of inertia into the rotational law formula: That's all.

Images