CN-121973738-A - Two-wheel vehicle ABS control system and method for estimating vision auxiliary attachment coefficient

CN121973738ACN 121973738 ACN121973738 ACN 121973738ACN-121973738-A

Abstract

The application relates to the technical field of active safety and braking control of two-wheel vehicles, in particular to a two-wheel vehicle ABS control system and method for estimating a visual auxiliary attachment coefficient.A visual priori generation module outputs visual priori information related to a road surface, a current attachment observation generation module outputs a longitudinal attachment observation value and a validity level, and the two information are sequentially fused with a closed loop updating module through attachment priori fusion, weight self-adaptive allocation and filtering treatment, and then the determined attachment coefficient is output to an ABS control system module to execute braking modulation; the fusion result checking module generates checking information, the closed-loop strategy state machine module outputs feedback instructions to adjust fusion and visual priori generation processes, stable estimation of attachment coefficients is achieved, ABS strategy switching is adapted, the problem of instability in single information source estimation is solved, and braking safety is improved.

Inventors

DENG SHUCHAO
WANG BO
TONG BAOHONG
JIAN ZHIQIANG
LIU PAN
WANG HUIRAN
ZHAO LINFENG

Assignees

安徽工业大学

Dates

Publication Date: 20260505
Application Date: 20260211

Claims (10)

1. The control method of the ABS of the two-wheel vehicle for estimating the vision auxiliary attachment coefficient is characterized by comprising the following steps of: Step one, data acquisition is synchronous, namely acquiring and synchronizing camera images, front and rear wheel speeds, inertial measurement unit data and master cylinder pressure data, and recording acquisition time stamps of the data; based on the collected camera image, performing image preprocessing, road surface identification and priori mapping, and outputting visual priori information comprising road surface category, attachment priori range, identification confidence and priori aging; thirdly, longitudinal observation generation, namely calculating longitudinal attachment observation values of front and rear wheels based on wheel speed, inertial measurement unit data and master cylinder pressure data, and outputting an observation effectiveness level according to master cylinder pressure grading and observation consistency indexes; Step four, single sequence fusion, namely carrying out input construction on visual priori and longitudinal observation, firstly judging whether available observation exists, executing a bottom-of-pocket output strategy when available observation is unavailable, executing update according to the sequence of 'first visual priori and then longitudinal observation', combining weight self-adaptive allocation and constraint filtering treatment, and outputting front and rear wheels to determine attachment coefficients; inputting the determined attachment coefficient into an ABS control module, selecting a corresponding control parameter set and a control algorithm branch, executing brake modulation, and outputting brake execution state and diagnosis information; step six, checking the fusion result, namely combining the fusion reliability, the brake execution state and the diagnosis information to generate check indexes, and outputting update permission, a conflict risk assessment result and strategy advice; step seven, closed-loop strategy feedback, namely based on a checking result, visual priori validity and longitudinal observation validity, performing gating judgment and state scheduling, outputting a closed-loop strategy instruction, feeding back to a fusion process and a visual priori generation link, and inhibiting output jump; And step eight, entering the next brake control period, and repeating the steps one to seven to realize the cyclic closed-loop control.
2. The ABS control method for the two-wheeled vehicle based on visual auxiliary attachment coefficient estimation according to claim 1 is characterized in that in the second step, visual priori generation comprises noise reduction and cutting pretreatment on road surface images acquired by a camera, a combined model of 'detection ROI+classification network' is adopted to identify road surface categories and output identification confidence, the range of the identification confidence is [0,1], a corresponding attachment priori range is output according to the road surface categories through a priori mapping table, and complete visual priori information is output after the priori timeliness of last effective identification of the distance is calculated.
3. The ABS control method for the two-wheeled vehicle based on the visual auxiliary attachment coefficient estimation according to claim 1, wherein in the third step, the observation effectiveness grade is classified into three grades of "effective", "weak effective" and "ineffective", when the master cylinder pressure is larger than or equal to a preset threshold value and the observation consistency deviation is smaller than or equal to a preset deviation threshold value, the judgment is made that the judgment is "weak effective", when only one of the master cylinder pressure and the consistency deviation meets a preset condition, the judgment is made that the judgment is "ineffective", and the judgment is made that the two items are distinguishable by longitudinal observation under the current working condition.
4. The method for controlling the ABS of the two-wheeled vehicle with visual auxiliary attachment coefficient estimation according to claim 1, wherein in the fourth step, the spam output strategy preferentially keeps the determined attachment coefficient output of the previous period, and if no history output data exists, the preset unknown road surface conservative attachment coefficient is output.
5. The method for controlling the ABS of the two-wheeled vehicle with the visual auxiliary attachment coefficient estimation according to claim 1, wherein in the fourth step, the weight self-adaptive distribution is based on recognition confidence, prior range width and prior aging construction visual false measurement uncertainty, the visual false measurement uncertainty is reduced along with the increase of the recognition confidence and is increased along with the increase of the prior range width and the prior aging, the longitudinal observation uncertainty is constructed on the basis of the observation effectiveness level, the higher the effectiveness is, the smaller the longitudinal observation uncertainty is, and the fusion weight is automatically distributed through the uncertainty of two paths of observation.
6. The method for controlling the ABS of the two-wheeled vehicle with visual auxiliary attachment coefficient estimation according to claim 1, wherein in the step seven, a stabilization mechanism of multi-period confirmation, hysteresis threshold and locking holding window is adopted for closed loop strategy feedback to distinguish transient conflict from continuous conflict, the continuous conflict is triggered to trigger re-identification, and weak update is executed on a working condition with higher risk.
7. The two-wheel vehicle ABS control system for estimating the visual auxiliary attachment coefficient is used for realizing the two-wheel vehicle ABS control method for estimating the visual auxiliary attachment coefficient according to any one of claims 1-6, and is characterized by comprising a visual priori generating module, a current attachment observation generating module, an attachment priori fusion and closed loop updating module, an ABS control system module, a fusion result checking module and a closed loop strategy state machine module; The visual priori generating module and the current attachment observation generating module are connected to the attachment priori fusion and closed loop updating module respectively, the attachment priori fusion and closed loop updating module is connected to the ABS control system module and the fusion result checking module respectively, the fusion result checking module is connected to the closed loop strategy state machine module, and the closed loop strategy state machine module is connected to the attachment priori fusion and closed loop updating module and the visual priori generating module respectively in a feedback mode.
8. The system for controlling the ABS of the two-wheeled vehicle with the visual auxiliary attachment coefficient estimation according to claim 7, wherein the visual priori generating module comprises a camera component, visual processing and calculating hardware, an image preprocessing module, a road surface recognition model and a priori mapping table, and the current attachment observation generating module comprises a front wheel speed sensor, a rear wheel speed sensor, an inertia measuring unit, a master cylinder pressure sensor, a vehicle speed estimating unit and a longitudinal observation constructing unit.
9. The system for controlling the ABS of the two-wheeled vehicle with visual auxiliary attachment coefficient estimation according to claim 7, wherein the attachment priori fusion and closed loop updating module comprises an uncertainty management unit, a sequential fusion updating unit, a weight self-adaptive distribution unit and a constraint filter processing unit, and when any observing input is unavailable, the sequential fusion updating unit skips the corresponding updating step, but does not terminate the whole fusion process, and the system is degraded.
10. The system of claim 7 wherein the closed loop policy state machine module comprises a state management unit, a judgment calculation unit and a policy action output unit, and the closed loop policy command comprises at least five states of normal fusion, prior hold, conflict observation, lock hold and re-identification wait, and the output closed loop policy command comprises at least an update strength adjustment command, a lock/unlock command and a request re-identification command.

Description

Two-wheel vehicle ABS control system and method for estimating vision auxiliary attachment coefficient Technical Field The application relates to the technical field of active safety and braking control of two-wheel vehicles, in particular to a two-wheel vehicle ABS control system and method for estimating a vision auxiliary attachment coefficient. Background The anti-lock control system for braking the two-wheel vehicle enables wheels to be maintained in a reasonable sliding range in the braking process by adjusting braking pressure, so that braking efficiency and vehicle stability are both considered. The road surface adhesion conditions directly determine the upper limit of braking force available to the wheels and the optimal slip target, thereby affecting control parameters, control strategies and braking strength in the anti-lock system. Compared with a four-wheel vehicle, the two-wheel vehicle is more susceptible to factors such as attitude change, load transfer, road surface fluctuation and the like during braking, and the attaching conditions of the front wheel and the rear wheel possibly have differences, so that higher requirements are put on the identification of the attaching conditions and the matching of control strategies. In the prior art, one type of method adopts signals such as wheel speed, vehicle longitudinal acceleration, vehicle body attitude information and the like, and constructs observed quantity through vehicle longitudinal dynamics relation and estimates the road surface adhesion level or can use the adhesion upper limit. Another method uses a camera and other visual sensors to identify the road surface material or wet and slippery state, outputs road surface category, and obtains the adhesion prior range or adhesion grade according to experience mapping to assist control strategy selection. Under the control scene of the current two-wheel vehicle brake actuating mechanism, the switching of control parameters or algorithm branches needs to be compatible with the real-time performance and the stability, namely, if the adhesion estimation output fluctuation is large or frequent jump is caused, frequent switching of the control parameters, high-frequency modulation of an actuator, reduced brake comfort and even stability risks can be caused, and if the adhesion switching is too conservative or lagged, the brake performance can be reduced and the brake distance can be prolonged. Therefore, how to obtain reliable and available attachment determination values for algorithm switching in the presence of uncertainty of different information sources is a technical problem still to be solved. In summary, the prior art at least faces the following challenges: 1. The requirement on the identifiability of the brake excitation and the working condition is high only by the attachment estimation of the longitudinal movement response, and the output reliability is insufficient in the initial stage of braking or when the observation is insufficient; 2. The attachment mapping is easily affected by the environment and imaging conditions only by relying on visual pavement recognition, and the mapping result is difficult to reflect the real-time change of the current available attachment; 3. Lack of an engineering method for fusing and weighting visual priori information and longitudinal observation information under uncertainty constraint and outputting a determined attachment coefficient which can be directly adopted by an ABS control module; 4. The lack of a supervision mechanism and an output stabilization means for the reliability of the fusion result makes it difficult to realize safe and controllable attachment switching when the road surface is suddenly changed, the error is recognized or the abnormality is observed. In view of the above, it is necessary to provide a two-wheel vehicle ABS control system and method for visually assisted attachment coefficient estimation, so as to achieve more reliable attachment determination and control policy switching under different road conditions, and improve the adaptability and safety of brake control. Disclosure of Invention The invention aims to solve the problems that in the prior art, when the adhesion coefficient is estimated only by relying on longitudinal motion information, the estimation is unstable due to insufficient braking excitation and unrecognizable influence of noise and working conditions, and the control parameter switching is unreliable due to the influence of misidentification and aging change on the current available adhesion when relying on visual identification. The utility model provides a two-wheeled vehicle ABS control system of vision auxiliary adhesion coefficient estimation, including the following steps: Step one, data acquisition is synchronous, namely acquiring and synchronizing camera images, front and rear wheel speeds, inertial measurement unit data and master cylinder pressure data, and recording acquisition time st