EPB anti-lock control method and system based on self-adaptive wheel speed filtering algorithm

Abstract

The invention discloses an EPB anti-lock control method and system based on a self-adaptive wheel speed filtering algorithm, and belongs to the field of anti-lock control, and the method comprises the following steps: S1, collecting driving parameters in a vehicle driving process in real time, extracting feature vectors in the driving parameters after preprocessing, inputting the feature vectors into a pre-trained driving condition classification model, and obtaining a driving condition classification model; outputting the current driving condition type; s2, correcting a wheel speed signal based on an extended Kalman filter with determined parameters; s3, calculating a tire pressure deviation ratio based on the corrected wheel speed signal; and S4, when it is detected that the tire pressure deviation ratio exceeds a set threshold value, an EPB anti-lock control system is triggered. By the adoption of the EPB anti-lock control method and system based on the self-adaptive wheel speed filtering algorithm, the wheel speed information can be accurately extracted under the complex road and environment conditions, the calculation precision of the tire pressure deviation ratio is remarkably improved, and therefore a more reliable basis is provided for starting emergency braking.

Description

Technical Field

[0001] The present invention relates to the technical field of anti-lock control, and particularly to an EPB anti-lock control method and system based on an adaptive wheel speed filtering algorithm. Background Art

[0002] In the process of the rapid development of the global automotive industry towards intelligence and electrification, vehicle active safety technology has become the core R & D direction to ensure road traffic safety. As the core technology for vehicle dynamic stability control, the Electronic Stability Program (ESP), its key execution unit, the Electronic Parking Brake (EPB), not only undertakes the static parking brake function, but also adjusts the wheel braking force in real time through the anti-lock control module in emergency braking scenarios to avoid out-of-control risks such as skidding and fishtailing caused by wheel locking. Specifically, the core control logic of the EPB anti-lock function relies on accurate wheel speed signals: that is, by collecting the rotational speeds of the four wheels in real time and calculating the tire pressure deviation ratio, the system can dynamically determine whether the wheels enter the critical state of locking, and then trigger the braking force adaptive adjustment strategy.

[0003] However, the anti-lock control accuracy of existing EPB systems still faces severe challenges under complex working conditions. Traditional wheel speed filtering algorithms generally use low-pass filters with fixed parameters (such as Kalman filtering, mean filtering), and their filtering coefficients cannot be dynamically adjusted according to the driving conditions, resulting in signal processing failure problems in the following typical scenarios: 1) When driving at high speed (>80 km / h), the noise signals (>50 Hz) generated by the wheel speed sensor due to the high-frequency excitation of road joints and tire tread patterns cannot be effectively filtered, resulting in periodic fluctuations of ±2 km / h in the wheel speed data; 2) When driving on a bumpy road at low speed (<20 km / h, road unevenness >8 mm), the instantaneous slip / slip signal of the wheel caused by non-uniform road excitation is coupled with the real wheel speed change, and the fixed filtering parameters result in a signal delay of more than 150 ms, causing a lag in the calculation of the tire pressure deviation ratio; 3) When braking emergently on a wet and slippery road (adhesion coefficient <0.3), the dynamic slip ratio of the wheel changes rapidly (0-30% change time <50 ms), and the dynamic response bandwidth of the traditional filter is insufficient (<10 Hz), unable to capture the high-frequency transient characteristics of the wheel speed, resulting in misjudgment of the anti-lock control threshold.

[0004] The above technical bottlenecks directly lead to a 15%-20% decrease in the calculation accuracy of the tire pressure deviation ratio, which may cause mis-triggering of the anti-lock system (false alarm rate increases by 30%) or response delay (braking pressure build-up time extends by 80 ms) in extreme working conditions, seriously threatening the safety of vehicle emergency braking. Summary of the Invention

[0005] The objective of the present invention is to provide an EPB anti-lock control method and system based on an adaptive wheel speed filtering algorithm to solve the above technical problems.

[0006] To achieve the above objective, the present invention provides an EPB anti-lock control method based on an adaptive wheel speed filtering algorithm, including the following steps: S1. Real-time collect the driving parameters during the vehicle driving process, extract the feature vectors from the driving parameters after preprocessing, input the feature vectors into a pre-trained driving condition classification model, and output the current driving condition type; S2. Determine the parameters of the extended Kalman filter according to the driving condition type, and correct the wheel speed signal based on the extended Kalman filter with the determined parameters; S3. Calculate the tire pressure deviation ratio based on the corrected wheel speed signal; S4. When it is detected that the tire pressure deviation ratio exceeds the set threshold, trigger the EPB anti-lock control system to perform a braking operation.

[0007] Preferably, the pre-trained driving condition classification model in step S1 is trained and tested with historical driving parameters, and during the training and testing process, the input of the driving condition classification model , respectively represent the feature vectors of the th sampling point, and the output , of the driving condition classification model respectively represent the driving condition types of the th sampling point; The driving parameters include the triaxial acceleration , the rotation angle , the tilt angle , the wheel speed and the tire pressure data to form the feature vectors.

[0008] Preferably, in step S1, the driving condition classification model is trained based on the random forest algorithm, which specifically includes the following steps: First, perform parameter initialization: determine the number of decision trees in the random forest and the maximum depth of each decision tree; Then construct a single decision tree: use the CART algorithm to construct the decision tree, and for each decision tree in the random forest, randomly select some features from the set of feature vectors of the original data, and at the same time randomly select some samples from the sample set for training; Finally, perform integrated decision-making: combine the prediction results of all decision trees, and use the majority voting method to determine the final driving condition classification result.

[0009] Preferably, step S2 specifically includes the following steps: S21. Customize the process noise adjustment parameter and the observation noise adjustment parameter of the driving condition type and the extended Kalman filter and the control strategy, and determine the current process noise adjustment parameter of the extended Kalman filter and the observation noise adjustment parameter according to the real-time driving condition type; S22. Correct the wheel speed signal based on the extended Kalman filter with the determined parameters: (1); (2); In the formula, and respectively represent the estimated values of the corrected wheel speed at moment and ; represents the control input; and respectively represent the process noise and the observation noise; and respectively represent the state transition function and the observation function; represents the collected wheel speed value; Use the Kalman gain to and for fusion to obtain the corrected wheel speed : (3); (4); (5); (6); In the formula, represents the error covariance when predicting the value at moment based on the estimated value at ; represents the error covariance at moment after correction; represents the error covariance after correction at moment; , , and respectively represent the state transition matrix, the process noise covariance matrix, the observation matrix, and the observation noise covariance matrix; represents the corrected wheel speed estimated value at moment; ​respectively represent the prediction based on the estimated value at the moment and the wheel speed value at the moment; represents the theoretical observed value obtained based on ; S23. Use formulas (1)-(6) to respectively obtain the corrected left wheel speed and the right wheel speed .

[0010] Preferably, the tire pressure deviation ratio described in step S3 is calculated as follows: (7).

[0011] Preferably, in step S4, when the tire pressure deviation ratio is lower than 80% of the set threshold, an alarm is triggered.

[0012] A system for implementing the EPB anti-lock control method based on the adaptive wheel speed filtering algorithm includes: A data acquisition and preprocessing module, configured to collect driving parameters during the vehicle driving process in real time, and extract feature vectors in the driving parameters after preprocessing; A driving condition classification model, configured to output the current driving condition type based on the feature vectors; A filtering parameter determination module, configured to determine the parameters of the extended Kalman filter according to the driving condition type, and correct the wheel speed signal based on the extended Kalman filter with the determined parameters; A tire pressure deviation ratio calculation module, configured to calculate the tire pressure deviation ratio based on the corrected wheel speed signal; A braking execution module, configured to trigger the EPB anti-lock control system and execute a braking operation when it detects that the tire pressure deviation ratio exceeds the set threshold.

[0013] Therefore, the present invention adopts the above-mentioned EPB anti-lock control method and system based on the adaptive wheel speed filtering algorithm, and the beneficial effects are as follows: 1. Accurate identification of multiple working conditions. Through the fusion of multi-source data such as acceleration sensors and vehicle body attitude sensors, combined with machine learning models (such as random forests), the driving conditions are identified in real time, and the accuracy of working condition identification reaches more than 95%, providing accurate input for the dynamic adjustment of subsequent filtering parameters, and solving the adaptability problem of the traditional fixed-parameter filtering algorithm's "undifferentiated processing" in complex working conditions; 2. The accuracy of the wheel speed signal is significantly improved, taking into account both noise suppression and dynamic response. Based on the working condition recognition results, the process noise and observation noise covariance matrices of the extended Kalman filter are dynamically adjusted. High-frequency noise above 50 Hz is effectively filtered out under high-speed working conditions. The signal delay is shortened to within 50 ms during low-speed bumpy driving. The dynamic response bandwidth of the wheel speed signal is increased to 20 Hz during emergency braking on a slippery road surface. The wheel speed measurement accuracy is improved by 20% compared with the traditional method, ensuring that the wheel speed signal is both smooth and can capture instantaneous changes. 3. The reliability of the anti-lock control is enhanced to ensure driving safety. The high-precision wheel speed signal reduces the calculation error of the tire pressure deviation ratio to less than 5%, avoiding calculation deviations caused by wheel speed signal noise. Under extreme working conditions, the mis-trigger rate of the anti-lock system is reduced by 40%, and the braking pressure build-up time is shortened by 60 ms. The EPB anti-lock control system can more accurately judge the risk of wheel lock-up and execute braking adjustment in a timely and reliable manner, significantly improving the stability and safety of the vehicle during emergency braking.

[0014] In summary, the present invention can dynamically optimize the filtering parameters according to the real-time driving conditions to solve the essential defect of traditional fixed-parameter filtering in multi-scenario adaptability, providing a high-precision wheel speed signal input for EPB anti-lock control. It not only solves the limitations of traditional fixed-parameter filtering algorithms in complex road conditions but also greatly improves the safety performance of the vehicle in emergencies, laying a foundation for building a safer and more efficient road traffic environment.

[0015] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Description of the Drawings

[0016] Figure 1 It is a flowchart of an EPB anti-lock control method based on an adaptive wheel speed filtering algorithm of the present invention. Detailed Embodiments

[0017] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the following further describes the embodiments of the present invention in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the embodiments of the present invention and are not used to limit the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts belong to the scope of protection of this application. The examples of the embodiments are shown in the accompanying drawings, where the same or similar reference numerals represent the same or similar elements or elements with the same or similar functions throughout.

[0018] It should be noted that the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units need not be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or devices.

[0019] The embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings.

[0020] As Figure 1 shown, an EPB anti-lock control method based on an adaptive wheel speed filtering algorithm includes the following steps: S1. Real-time collect the driving parameters during the vehicle driving process, extract the feature vectors from the driving parameters after preprocessing, and input the feature vectors into a pre-trained driving condition classification model to output the current driving condition type; the driving conditions in this embodiment include high-speed driving, low-speed bumpy roads, slippery roads, etc.

[0021] The pre-trained driving condition classification model described in step S1 is trained and tested with historical driving parameters, and during the training and testing process, the input of the driving condition classification model is set , respectively represent the feature vectors of the th sampling point, and the output of the driving condition classification model, respectively represent the driving condition types of the th sampling point; The driving parameters include triaxial acceleration , rotation angle , tilt angle , wheel speed and tire pressure data to form feature vectors.

[0022] In step S1, the driving condition classification model is trained based on the random forest algorithm, which specifically includes the following steps: First, perform parameter initialization: determine the number of decision trees in the random forest and the maximum depth of each decision tree; Then construct a single decision tree: use the CART algorithm to construct the decision tree, and for each decision tree in the random forest, randomly select some features from the set of feature vectors of the original data, and at the same time randomly select some samples from the sample set for training; Finally, perform integrated decision-making: combine the prediction results of all decision trees and use the majority voting method to determine the final driving condition classification result.

[0023] In this embodiment, cross-validation or independent test sets are also used to evaluate model performance. The evaluation indicators include accuracy, precision, recall, and F1-Score. Driving condition classification: assuming that the driving conditions are divided into three conditions: high-speed driving, low-speed bumpy road, and slippery road. Representation. Based on the trained machine learning model, new sensor data can be classified in real time.

[0024] S2. Determine the parameters of the extended Kalman filter according to the driving condition type, and correct the wheel speed signal based on the extended Kalman filter with the determined parameters; Step S2 specifically includes the following steps: S21, Customized driving condition type and process noise adjustment parameters of extended Kalman filter and observation noise adjustment parameter Control strategy, and determine the process noise adjustment parameters of the current extended Kalman filter according to the real-time driving condition type and observation noise adjustment parameter ; and Participate in controlling the filter to adapt to different dynamic environments; Represents a parameter related to the filter gain, which determines how much the filter trusts new measurement data. A higher value means the filter is more likely to accept new measurements, while a lower The value indicates that the filter is more dependent on the state predicted by its internal model. Therefore, the following control strategy can be customized according to the actual situation, for example: High speed driving: In this case, since the vehicle is moving more smoothly, a smaller and , reducing sensitivity to short-term fluctuations, thereby obtaining a smoother speed estimate. Slow-speed bumpy or slippery roads: At this time, due to the presence of more instantaneous interference, the and , in order to quickly respond to changes and reduce the probability of misjudgment caused by road conditions.

[0025] S22, correcting the wheel speed signal based on the extended Kalman filter with determined parameters: (1); (2); In the formula, and Respectively represent the corrected wheel speed in Moment and Estimated value; Represents the control input; And Respectively represent the process noise and the observation noise; And Respectively represent the state transition function and the observation function; Represents the collected wheel speed value; Using the Kalman gain To And Are fused to obtain the corrected wheel speed : (3); (4); (5); (6); In the formula, Represents the error covariance when predicting the predicted Time value based on the estimated value at Time; Represents the error covariance at Time after correction; Represents The error covariance after correction at , , And Respectively represent the state transition matrix, the process noise covariance matrix, the observation matrix, and the observation noise covariance matrix; Represents The corrected wheel speed estimated value at Respectively represent predicting Based on the estimated value at The wheel speed value at Represents The theoretical observation value obtained based on S23. Using formulas (1) - (6), the corrected left wheel speed And the right wheel speed Are obtained.

[0026] S3. Based on the corrected wheel speed signal, calculate the tire pressure deviation ratio; The tire pressure deviation ratio described in step S3 The calculation formula is as follows: (7).

[0027] S4. When it is detected that the tire pressure deviation ratio exceeds the set threshold, trigger the EPB anti-lock control system and perform the braking operation.

[0028] In step S4, when the tire pressure deviation ratio is lower than 80% of the set threshold, an alarm is triggered.

[0029] A system for implementing an EPB anti-lock control method based on an adaptive wheel speed filtering algorithm, comprising: A data acquisition and preprocessing module, configured to collect driving parameters during vehicle driving in real time, and extract feature vectors from the driving parameters after preprocessing; A driving condition classification model, configured to output the current driving condition type based on the feature vectors; A filtering parameter determination module, configured to determine the parameters of an extended Kalman filter according to the driving condition type, and correct the wheel speed signal based on the extended Kalman filter with the determined parameters; A tire pressure deviation ratio calculation module, configured to calculate the tire pressure deviation ratio based on the corrected wheel speed signal; A braking execution module, configured to trigger an EPB anti-lock control system and execute a braking operation when it detects that the tire pressure deviation ratio exceeds the set threshold.

[0030] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that they can still modify or equivalently replace the technical solutions of the present invention, and these modifications or equivalent replacements do not make the modified technical solutions deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. An EPB anti-lock control method based on an adaptive wheel speed filtering algorithm, characterized in that: It includes the following steps: S1. Collect the driving parameters during the vehicle driving in real time, extract the feature vectors in the driving parameters after preprocessing, input the feature vectors into the pre-trained driving condition classification model, and output the current driving condition type; S2. Determine the parameters of the extended Kalman filter according to the driving condition type, and correct the wheel speed signal based on the extended Kalman filter with the determined parameters; S3. Calculate the tire pressure deviation ratio based on the corrected wheel speed signal; S4. When it is detected that the tire pressure deviation ratio exceeds the set threshold, trigger the EPB anti-lock control system and perform the braking operation.

2. The EPB anti-lock control method based on the adaptive wheel speed filtering algorithm according to claim 1, characterized in that: The pre-trained driving condition classification model described in step S1 is trained and tested with historical driving parameters, and during the training and testing processes, the input of the driving condition classification model is set , respectively represent the feature vectors of the th sampling point, and the output of the driving condition classification model , respectively represent the driving condition types of the th sampling point; The driving parameters include three-axis acceleration , rotation angle , tilt angle , wheel speed and tire pressure data , forming a feature vector.

3. The EPB anti-lock control method based on the adaptive wheel speed filtering algorithm according to claim 2, wherein: In step S1, the driving condition classification model is trained based on the random forest algorithm, which specifically includes the following steps: First, perform parameter initialization: determine the number of decision trees in the random forest and the maximum depth of each decision tree ; Then construct a single decision tree: Use the CART algorithm to construct the decision tree, and for each decision tree in the random forest, randomly select some features from the set of feature vectors of the original data, and at the same time randomly select some samples from the sample set for training; Finally, integrate the decisions: Combine the prediction results of all decision trees and use the majority voting method to determine the final driving condition classification result.

4. The EPB anti-lock control method based on the adaptive wheel speed filtering algorithm according to claim 2, wherein: Step S2 specifically includes the following steps: S21. Process of customizing driving condition types and adjusting parameters of process noise and observation noise of the extended Kalman filter and the observation noise adjustment parameter control strategy, and determine the current process noise adjustment parameter of the extended Kalman filter according to the real-time driving condition type and the observation noise adjustment parameter ; S22. Correct the wheel speed signal based on the extended Kalman filter with the determined parameters: （1）； （2）； Wherein, and respectively represent the estimated values of the corrected wheel speed at moment and ; represents the control input; and respectively represent the process noise and the observation noise; and respectively represent the state transition function and the observation function; represents the collected wheel speed value; Using the Kalman gain Pair and are fused to obtain the corrected wheel speed : （3）； （4）； （5）； （6）； In the formula, represents the error covariance when predicting the predicted time value based on the estimated value at time; represents the error covariance at time after correction; represents the error covariance after correction at time; , , and respectively represent the state transition matrix, the process noise covariance matrix, the observation matrix, and the observation noise covariance matrix; represents the estimated wheel speed value after correction at respectively represent the predicted wheel speed value at time based on the estimated value at represents the theoretical observation value obtained based on ; S23. Respectively obtain the corrected left wheel speed and right wheel speed by using Formula (1) - Formula (6). .​ 5. The EPB anti-lock control method based on the adaptive wheel speed filtering algorithm according to claim 4, characterized in that: The tire pressure deviation ratio described in step S3 The calculation formula is as follows: （7）。 6. The EPB anti-lock control method based on the adaptive wheel speed filtering algorithm according to claim 5, characterized in that: In step S4, when the tire pressure deviation ratio is lower than 80% of the set threshold, an alarm is triggered.

7. A system for implementing the EPB anti-lock control method based on the adaptive wheel speed filtering algorithm according to any one of claims 1-6 above, characterized in that: It includes: A data collection and preprocessing module, which is used to collect the driving parameters during the vehicle driving in real time and extract the feature vectors in the driving parameters after preprocessing; A driving condition classification model, which is used to output the current driving condition type based on the feature vectors; A filtering parameter determination module, which is used to determine the parameters of the extended Kalman filter according to the driving condition type and correct the wheel speed signal based on the extended Kalman filter with the determined parameters; A tire pressure deviation ratio calculation module, which is used to calculate the tire pressure deviation ratio based on the corrected wheel speed signal; A braking execution module, which is used to trigger the EPB anti-lock control system and perform the braking operation when it is detected that the tire pressure deviation ratio exceeds the set threshold.

Images