Real-time fault detection method for vehicle electronic braking system

Abstract

The invention relates to the field of vehicle braking fault detection, in particular to a real-time fault detection method for a vehicle electronic braking system, which comprises the following steps of: acquiring vehicle driving state data through a vehicle-mounted sensor network and an electronic control unit, calculating filtered residual data of a target wheel speed sensor, and calculating the filtered residual data of the target wheel speed sensor; establishing calibration reference parameters of the health state; performing conjoint analysis on the acceleration data and the residual fluctuation data of the vehicle to obtain a dynamic response adaptive adjustment factor of the vehicle; obtaining a tiny drift confirmation factor of the vehicle through the candidate residual error in the low dynamic interference state of the vehicle; continuous fault judgment is performed by combining the dynamic response adjustment factor and the tiny drift confirmation factor of the vehicle, and the fault judgment result of the wheel speed sensor is obtained, so that the accuracy and sensitivity of fault detection when tiny systematic deviation occurs to the wheel speed sensor are improved.

Description

Technical Field

[0001] The present invention relates to the field of vehicle braking fault detection, and particularly to a real-time fault detection method for a vehicle electronic braking system. Background Art

[0002] Modern vehicles are generally equipped with electronic braking systems, such as an anti-lock braking system, an electronic stability control program, and a traction control system, etc. These systems significantly enhance the braking performance, driving stability, and overall safety of the vehicle under complex driving conditions by precisely regulating the braking force of the braking system or the driving force of the vehicle. As a core sensing component of the electronic braking system, the wheel speed sensor is responsible for measuring the rotational speed of each wheel in real time and accurately, providing the basic data necessary for the electronic control unit to make decisions and controls. Based on the signals provided by the wheel speed sensor, the electronic control unit judges the immediate motion state of the wheels, such as whether wheel lock-up occurs, whether the driving wheel slips, or whether the vehicle shows an instability trend, and accordingly quickly adjusts the braking pressure or the output torque of the engine. Therefore, the accuracy and stability of the wheel speed sensor signals play a decisive role in whether the electronic braking system can effectively perform its safety guarantee duty. Currently, for the fault diagnosis of the wheel speed sensor, the commonly used technical means is the method based on model comparison. This method estimates the expected wheel speed of the target wheel by constructing a reference model of the vehicle and using information such as the wheel speeds of other wheels and vehicle dynamic parameters. Then, the rotational speed data actually output by the wheel speed sensor is compared with the expected rotational speed predicted by this model, thereby obtaining a residual data. In the traditional fault judgment logic, when the amplitude of this residual data continuously exceeds a preset fixed threshold within a period of time, the system determines that the corresponding wheel speed sensor has failed.

[0003] However, in the above prior art, the method of fault judgment based on a fixed threshold for the residual data of the wheel speed sensor has significant deficiencies in the early weak fault identification method of the wheel speed sensor. Specifically, when the wheel speed sensor undergoes a gradual change in its own performance or a change in the working environment, the data it outputs shows a systematic deviation with a small amplitude and intermittent occurrence, that is, the wheel speed reading continuously slightly higher or lower than the true wheel speed of the vehicle. Such a small systematic deviation, as an early sign of potential performance degradation of the sensor, will gradually develop into a more significant fault, and is more likely to have an adverse impact on the control accuracy and response timeliness of the system at the critical moment when functions such as anti-lock braking or electronic stability control system in the electronic braking system are accessed, thus threatening the driving safety and stability of the vehicle. The main limitation of the traditional fixed threshold method is that in order to avoid false alarms during normal dynamic driving of the vehicle, its threshold setting is usually relatively loose. This loose threshold directly leads to insensitivity to the above-mentioned small systematic deviation, making the residual signal generated by this deviation unable to trigger an alarm or be misidentified as normal signal noise due to its too small amplitude, resulting in missed detection of faults. Therefore, how to accurately and sensitively detect this intermittent, small-amplitude but unidirectional persistent systematic deviation of the wheel speed sensor under the premise of effectively adapting to the normal dynamic changes of the vehicle and suppressing noise interference has become a key technical problem urgently to be solved in the field of vehicle electronic braking system fault detection. Summary of the Invention

[0004] In view of this, an embodiment of the present invention provides a real-time fault detection method for a vehicle electronic braking system to solve the problems of low accuracy and poor sensitivity in fault detection when a small systematic deviation appears in the wheel speed sensor.

[0005] An embodiment of the present invention provides a real-time fault detection method for a vehicle electronic braking system, and the method includes the following steps: Step S1: Collect vehicle driving state data through an in-vehicle sensor network and an electronic control unit, calculate the filtered residual data of the target wheel speed sensor, and obtain the calibration reference parameters of the healthy state; Step S2: Obtain the dynamic response adaptive adjustment factor of the vehicle by jointly analyzing the vehicle dynamic energy and the volatility of the residual signal through the vehicle acceleration data and the residual fluctuation data; Step S3: Obtain the small drift confirmation factor of the vehicle by performing a significance analysis on the consistency and amplitude of the candidate residuals under the low dynamic interference state of the vehicle; Step S4: Obtain the fault determination result of the small systematic deviation of the wheel speed sensor by combining the dynamic response adjustment factor and the small drift confirmation factor of the vehicle for continuous fault judgment; Step S5: Complete the active maintenance response of the wheel speed sensor by performing signal reporting and status storage processing on the fault determination result.

[0006] Preferably, the method of collecting vehicle driving state data through the vehicle sensor network and the electronic control unit, calculating the filtered residual data of the target wheel speed sensor, and establishing the calibration reference parameters for the health state includes: Set the data sampling period, and collect vehicle driving state data through the vehicle sensor network and the electronic control unit. The collected data includes: the longitudinal acceleration of the vehicle, the lateral acceleration of the vehicle, the vehicle reference speed, and the original wheel speed data output by the vehicle target wheel speed sensor; Obtain the expected wheel speed of the vehicle target wheel speed sensor through the vehicle reference speed and the dynamic tire radius of the target wheel; take the difference between the original wheel speed data and the expected wheel speed as the original residual data of the vehicle target wheel speed; perform filtering processing on the original residual data of the vehicle target wheel speed through a second-order Butterworth low-pass filter to obtain the residual data of the vehicle target wheel speed; Obtain the reference dynamic energy data of the normal dynamic energy level of the vehicle, the reference residual fluctuation data of the inherent fluctuation level of the residual signal when the wheel speed sensor is working normally, and the baseline noise standard deviation data of the residual data in the healthy state under the condition of steady-speed driving through the offline calibration process of the vehicle; store the reference dynamic energy data, the reference residual fluctuation data, and the baseline noise standard deviation data as reference parameters in the electronic control unit of the vehicle.

[0007] Preferably, the method of obtaining the dynamic response adaptive adjustment factor of the vehicle by jointly analyzing the vehicle acceleration data and the residual fluctuation data includes: Perform vehicle dynamic energy analysis on the longitudinal acceleration and lateral acceleration of the vehicle to obtain the first vehicle dynamic energy index at the target moment; Set the time series window of the wheel speed sensor residual data, and obtain the short-term wheel speed residual standard deviation of the wheel speed sensor residual data in the previous time series window at the target moment; Obtain the dynamic response adaptive adjustment factor of the vehicle by performing a fusion evaluation of the first vehicle dynamic energy index and the short-term wheel speed residual standard deviation on the dynamic driving intensity and the wheel speed residual fluctuation level.

[0008] Preferably, the method of performing vehicle dynamic energy analysis on the longitudinal acceleration and lateral acceleration of the vehicle to obtain the first vehicle dynamic energy index at the target moment includes: Set the length of the sliding time window for dynamic energy analysis; obtain the sum of the squares of the lateral acceleration and the longitudinal acceleration of the vehicle at each timestamp in the sliding time window for dynamic energy analysis; take the square root of the mean of the sum of the squares of the accelerations corresponding to all timestamps in the sliding time window as the first vehicle dynamic energy index at the target moment.

[0009] Preferably, the method for obtaining the dynamic response adaptive adjustment factor of the vehicle by comprehensively evaluating the dynamic driving intensity and the wheel speed residual fluctuation level through the first vehicle dynamic energy index and the short-term wheel speed residual standard deviation includes: Obtain the reference dynamic energy data and the reference residual fluctuation data; take the ratio of the first vehicle dynamic energy index to the reference dynamic energy data as the first adjustment factor; take the ratio of the short-term wheel speed residual standard deviation to the reference residual fluctuation data as the second adjustment factor; take the calculation result of adding the constant 1, the first adjustment factor and the second adjustment factor as the dynamic response adaptive adjustment factor of the vehicle.

[0010] Preferably, the method for obtaining the micro drift confirmation factor of the vehicle by significantly analyzing the consistency and amplitude of the candidate residuals under the low dynamic interference state of the vehicle includes: Set the basic threshold of the vehicle wheel speed residual data, and optimize the basic threshold by weighting with the dynamic response adaptive adjustment factor of the vehicle to obtain the first dynamic threshold of the vehicle wheel speed residual data; Preliminarily screen the vehicle wheel speed residual data through the first dynamic threshold of the vehicle wheel speed residual data to obtain the candidate residual data of the vehicle wheel speed; Obtain the conditional cumulative consistency score by evaluating the consistency of the candidate residual data; obtain the relative degree score of the candidate residual data condition amplitude exceeding the baseline noise by analyzing the amplitude change of the candidate residual data; Obtain the micro drift confirmation factor of the vehicle through the comprehensive evaluation of the conditional cumulative consistency score and the relative degree score of the candidate residual data condition amplitude exceeding the baseline noise.

[0011] Preferably, the method for obtaining the conditional cumulative consistency score of the candidate residual data by evaluating the consistency of the candidate residual data and obtaining the relative degree score of the candidate residual data condition amplitude exceeding the baseline noise by analyzing the amplitude change of the candidate residual data includes: Set the length of the long observation time window for judging minute drift; divide the window of vehicle wheel speed residual data through the length of the long observation time window to obtain the first window of vehicle wheel speed residual data; evaluate the driving state through the real-time driving speed, lateral acceleration and longitudinal acceleration of the vehicle to obtain the evaluation result of the vehicle steady speed driving condition. If the vehicle is in the steady speed driving state, the evaluation result of the vehicle steady speed driving condition is 1; if the vehicle is not in the steady speed driving state, the evaluation result of the vehicle steady speed driving condition is 0. Take the candidate residual data that meets the evaluation result of the vehicle steady speed driving condition being 1 in the first window of the vehicle wheel speed residual data at the target moment as the second candidate wheel speed residual data at the target moment; take the absolute value of the sum of all the second candidate wheel speed residual data in the first window at the target moment as the cumulative deviation direction evaluation at the target moment; take the summation calculation result of the absolute values of all the second candidate wheel speed residual data in the first window at the target moment as the total fluctuation magnitude evaluation at the target moment; take the ratio of the cumulative deviation direction evaluation at the target moment to the total fluctuation magnitude evaluation at the target moment as the conditional cumulative consistency score of the candidate residual data at the target moment. Set the significance factor of the candidate residual; take the calculation result of multiplying the baseline noise standard deviation data by the significance factor as the significance standard deviation of the candidate residual; take the calculation result of subtracting the mean value of the absolute values of the second candidate wheel speed residual data at the target moment from the significance standard deviation and dividing by the baseline noise standard deviation data as the first relative degree evaluation at the target moment; take the larger value between the first relative degree evaluation at the target moment and the constant 0 as the relative degree score of the candidate residual data condition amplitude exceeding the baseline noise.

[0012] Preferably, the comprehensive evaluation through the conditional cumulative consistency score and the relative degree score of the vehicle wheel speed residual data condition amplitude exceeding the baseline noise to obtain the minute drift confirmation factor of the vehicle includes: Take the calculation result of multiplying the conditional cumulative consistency score of the candidate residual data at the target moment by the relative degree score of the candidate residual data condition amplitude exceeding the baseline noise as the minute drift confirmation factor of the vehicle.

[0013] Preferably, the continuous fault judgment through combining the dynamic response adjustment factor of the vehicle and the minute drift confirmation factor to obtain the fault determination result of the minute systematic deviation of the wheel speed sensor includes: Set the fault judgment threshold and the length of the fault confirmation duration period, initialize the fault confirmation counter. During the data sampling period, if the micro-drift confirmation factor of the vehicle is greater than the fault judgment threshold, the fault confirmation counter is incremented; if the micro-drift confirmation factor of the vehicle is less than or equal to the fault judgment threshold, the fault confirmation counter is reset to 0; when the cumulative value of the fault confirmation counter reaches the duration period length, it is determined that the target wheel speed sensor has a micro-systematic deviation fault.

[0014] The beneficial effects of the embodiments of the present invention compared with the prior art are as follows: The present invention provides a real-time fault detection method for a vehicle electronic braking system. Compared with the existing diagnostic methods that only rely on fixed thresholds or instantaneous fluctuations for judgment, the present invention first establishes a joint analysis mechanism based on the difference between the vehicle's dynamic state and the response of the wheel speed sensor. The system constructs a residual signal representing the deviation degree between the vehicle operation and the wheel speed expectation by obtaining the vehicle's longitudinal and lateral accelerations, reference driving speed, and the original signals of the wheel speed sensor in real time, and suppresses high-frequency noise and occasional disturbances through signal processing techniques on this basis, so as to obtain more stable and discriminable signal features. This process ensures a stable input basis for the subsequent diagnostic model in terms of the quality of the original signal. In order to fully consider the rationality of the wheel speed signal fluctuations under different driving conditions and improve the system's adaptability to complex dynamic behaviors, the present invention introduces a calibration mechanism in the system initialization stage. By collecting and statistically modeling the vehicle operation data and wheel speed response results under a variety of typical driving conditions over a long period, benchmark values that can reflect the normal vehicle driving characteristics are obtained, including dynamic behavior reference values, signal fluctuation level benchmark values, and noise lower limits under static driving conditions. Through the construction of this calibration model, the system can adaptively adjust the detection sensitivity under different vehicles and different road conditions, significantly improving the generality and deployment efficiency of the algorithm. In the discrimination mechanism, the present invention first proposes a dual-factor diagnostic model in which a dynamic adjustment factor and a micro-drift confirmation factor work together. Among them, the design of the dynamic adjustment factor fully considers the influence of the actual vehicle operation state on the wheel speed signal fluctuations. When the vehicle is in a high-speed driving, rapid acceleration, emergency braking, or steering state, the initial residual recognition threshold is automatically relaxed, effectively preventing the system from misjudging the dynamic disturbances generated by the vehicle itself as sensor abnormalities; while when the vehicle is in a low-dynamic state such as stable cruising or uniform driving, the recognition threshold is automatically tightened, thereby enhancing the system's perception ability of weak but persistent abnormal deviations and achieving a dynamic balance between false alarm control in the dynamic state and abnormal detection sensitivity in the static state.

[0015] The present invention also introduces a continuous confirmation mechanism. After the diagnostic system determines a suspected deviation signal, it needs to go through consecutive judgment cycles and maintain consistency before the fault state can be confirmed. This mechanism significantly reduces the risk of misjudgment caused by non-structural events such as occasional vibrations and short-term electrical interferences, making the final diagnostic result more stable and reliable. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 is a flowchart of a real-time fault detection method for a vehicle electronic braking system provided in Embodiment 1 of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0017] Embodiments of the present disclosure will be described in detail below. Examples of the embodiments are shown in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to explain the present disclosure and should not be construed as limiting the present disclosure.

[0018] To illustrate the technical solution of the present invention, specific embodiments will be used for illustration below.

[0019] Refer to Figure 1 , which is a flowchart of a real-time fault detection method for a vehicle electronic braking system provided in Embodiment 1 of the present invention. As Figure 1 shown, the method may include: Step S1, collect vehicle driving state data through an in-vehicle sensor network and an electronic control unit, calculate the filtered residual data of the target wheel speed sensor, and obtain the calibration reference parameters of the healthy state.

[0020] First, during the operation of the vehicle, through the in-vehicle sensor network and the relevant electronic control unit, with a sampling period of 10 milliseconds, that is, a corresponding sampling frequency of 100 Hz, collect the data related to the vehicle driving state and the operation of the wheel speed sensor in real time. Specifically, the collected data includes the longitudinal acceleration of the vehicle, the lateral acceleration of the vehicle, the vehicle reference speed, and the original wheel speed data output by the vehicle target wheel speed sensor.

[0021] Based on the above data collection, further calculate the expected wheel speed of the target wheel speed sensor at the target moment by the vehicle reference speed and the dynamic tire radius of the target wheel. Specifically, calculate the expected wheel speed by dividing the vehicle reference speed by the dynamic tire radius of the target wheel.

[0022] After obtaining the desired wheel speed, the original wheel speed residual data at the target moment is calculated by subtracting the wheel speed actually measured by the target wheel speed sensor from the desired wheel speed. To eliminate the inherent high-frequency noise and irregular fluctuations in the original wheel speed residual data, and thus extract more stable and reliable fault characteristics, the original wheel speed residual data is filtered through a second-order Butterworth low-pass filter to obtain the filtered residual data, and the filtered residual data is used as the residual data of the vehicle's target wheel speed.

[0023] After obtaining the residual data of the vehicle's target wheel speed, through an offline calibration process, the reference parameters in the healthy state are obtained. Specifically, a calibration vehicle with the working state of the wheel speed sensor confirmed to be normal is selected, and under the test conditions covering various typical driving conditions, a large amount of the above-defined healthy state data is collected, especially the vehicle acceleration data and the wheel speed residual data after filtering. Based on this healthy state data set, the following calibration reference parameters are calculated and determined, including: the reference dynamic energy data of the vehicle's normal dynamic energy level, the reference residual fluctuation data of the inherent fluctuation level of the residual signal when the wheel speed sensor is working normally, and the baseline noise standard deviation data of the healthy state residual data under the condition of steady-speed driving; the reference dynamic energy data, the reference residual fluctuation data and the baseline noise standard deviation data are stored as reference parameters in the vehicle's electronic control unit.

[0024] Step S2, by jointly analyzing the vehicle acceleration data and the residual fluctuation data for the vehicle's dynamic energy and the volatility of the residual signal, the dynamic response adaptive adjustment factor of the vehicle is obtained.

[0025] During the actual operation of the vehicle, its driving state is complex and changeable, including various dynamic conditions such as acceleration, deceleration, steering, and passing through uneven roads. Under these dynamic conditions, even if the wheel speed sensor works perfectly normally, due to physical phenomena such as the interaction between the wheel and the road surface, the transfer of vehicle load, and the elastic deformation of the tire, the instantaneous rotational speeds of each wheel will produce reasonable or even large fluctuations. These normal wheel speed fluctuations will inevitably be reflected as corresponding fluctuation components in the residual signal when compared with the vehicle reference model. The prior art uses a fixed threshold to judge whether the residual signal is abnormal. Then, in order to avoid a large number of false fault alarms during these normal vehicle dynamic driving, this fixed threshold must be set relatively high. However, a relatively high fixed threshold will directly result in being insensitive to those systematic deviations with small amplitudes that indicate early faults or performance gradual changes of the sensor, thus leading to the missed detection of such real faults.

[0026] Therefore, to solve the inherent contradiction between the fixed threshold in adapting to vehicle dynamic changes and detecting minor faults, the present invention first introduces a vehicle dynamic response adaptive adjustment factor. This factor dynamically adjusts the preliminary benchmark for fault judgment based on the current real-time dynamic driving intensity of the vehicle and the inherent fluctuation level of the wheel speed sensor residual signal in the recent period. When the vehicle is in a severe dynamic working condition (such as sudden acceleration, sudden braking, or rapid steering), at this time, large-amplitude fluctuations in the wheel speed residual signal are normal phenomena. The vehicle dynamic response adaptive adjustment factor should be able to correspondingly increase the judgment benchmark to tolerate these reasonable dynamic disturbances and prevent false alarms. On the contrary, when the vehicle is in a steady driving state and the residual signal itself has small fluctuations, any minor residual deviation from the normal baseline is more worthy of attention. At this time, the vehicle dynamic response adaptive adjustment factor should make the judgment benchmark more stringent to improve the sensitivity to weak abnormal signals. In this way, the vehicle dynamic response adaptive adjustment factor enables the preliminary screening process of fault detection to intelligently adapt to the real-time operating state of the vehicle, providing a more reliable input for the subsequent more refined identification of minor systematic deviations.

[0027] For the vehicle dynamic response adaptive adjustment factor, first, through vehicle dynamic energy analysis of the vehicle longitudinal acceleration and lateral acceleration, obtain the first vehicle dynamic energy index at the target moment; then, set the time series window of the wheel speed sensor residual data, obtain the short-term wheel speed residual standard deviation of the wheel speed sensor residual data in the previous time series window at the target moment, and through the fusion evaluation of the first vehicle dynamic energy index and the short-term wheel speed residual standard deviation for the dynamic driving intensity and wheel speed residual fluctuation level, obtain the vehicle dynamic response adaptive adjustment factor.

[0028] Specifically, first, set the sliding time window length for dynamic energy analysis. In this embodiment, set the sliding time window length to the number of sampling points in 0.5 seconds; obtain the sum of the squares of the acceleration of the vehicle lateral acceleration and longitudinal acceleration at each time stamp in the sliding time window for dynamic energy analysis; take the square root of half of the mean of the sum of the squares of the acceleration corresponding to all time stamps in the sliding time window as the first vehicle dynamic energy index at the target moment.

[0029] In one embodiment, assume that the vehicle longitudinal acceleration at the th historical moment in the sliding window for dynamic energy analysis is ; the vehicle lateral acceleration at the th historical moment in the sliding window for dynamic energy analysis is ; the number of data points in the sliding window for dynamic energy analysis is , then the calculation formula for the first vehicle dynamic energy index at the th moment is:

[0030] Among them, represents the first vehicle dynamic energy index at the -th moment; represents the vehicle longitudinal acceleration at the -th historical moment; represents the vehicle lateral acceleration at the -th historical moment; represents the number of data points in the sliding window for dynamic energy analysis.

[0031] After obtaining the first vehicle dynamic energy index at the target moment, set the time series window of the wheel speed sensor residual data. In the embodiment of the present invention, the length of the time series window of the wheel speed sensor residual data is set to the sampling length corresponding to 3 seconds, and the short-term wheel speed residual standard deviation of the wheel speed sensor residual data in the previous time series window at the target moment is obtained.

[0032] In one embodiment, assume that the residual data of the vehicle target wheel speed at the -th historical moment is ; the mean value of the residual data in the previous time series window of the wheel speed sensor residual data at the -th target moment is ; the number of data points in the time series window of the wheel speed sensor residual data is , then the calculation formula for the short-term wheel speed residual standard deviation of the wheel speed sensor residual data in the previous time series window at the -th target moment is:

[0033] Among them, represents the short-term wheel speed residual standard deviation of the wheel speed sensor residual data in the previous time series window at the -th target moment; represents the number of data points in the time series window of the wheel speed sensor residual data; represents the residual data of the vehicle target wheel speed at the -th historical moment; represents the mean value of the residual data in the previous time series window of the wheel speed sensor residual data at the -th target moment.

[0034] Finally, obtain the reference dynamic energy data and the reference residual fluctuation data; take the ratio of the first vehicle dynamic energy index to the reference dynamic energy data as the first adjustment factor; take the ratio of the short-term wheel speed residual standard deviation to the reference residual fluctuation data as the second adjustment factor; take the calculation result of adding the constant 1, the first adjustment factor, and the second adjustment factor as the dynamic response adaptive adjustment factor of the vehicle. For the vehicle at the th target moment, the calculation formula of the dynamic response adaptive adjustment factor is:

[0035] where, represents the dynamic response adaptive adjustment factor of the vehicle at the th target moment; represents the first vehicle dynamic energy index at the th moment; represents the reference dynamic energy data of the normal dynamic energy level of the vehicle obtained through the off-line calibration process of the vehicle; represents the short-term wheel speed residual standard deviation of the wheel speed sensor residual data in the previous time window of the th target moment; represents the reference residual fluctuation data of the inherent fluctuation level of the residual data when the wheel speed sensor works normally, obtained through the off-line calibration process of the vehicle.

[0036] It should be noted that the dynamic response adaptive adjustment factor proposed by the present invention aims to overcome the inherent defect of poor adaptability of the traditional fixed threshold under complex vehicle dynamics; firstly, the constant 1 in the calculation formula of the dynamic response adaptive adjustment factor is the basic value set for this adjustment factor, indicating that in the extreme case of ideal vehicle being completely stationary and sensor data having no fluctuation, the value of the dynamic response adaptive adjustment factor of the vehicle approaches 1, and at this time, the dynamic effective threshold will be mainly determined by the preset basic threshold, so as to ensure that even under static conditions, the system still retains the basic detection ability for minor anomalies. Secondly, the second term in the calculation formula of the dynamic response adaptive adjustment factor is the compensation for the macroscopic dynamic driving intensity of the vehicle. During the driving process of the vehicle, its longitudinal acceleration and lateral acceleration are the key physical quantities characterizing its dynamic behavior. By calculating the root mean square of the sum of squares of these accelerations, the overall dynamic energy of the vehicle within the recent time window is effectively quantified. When the vehicle performs high-dynamic operations such as emergency acceleration or obstacle avoidance steering, will increase significantly. At this time, by dividing by the reference dynamic energy A relative ratio is obtained to reflect how many times the current vehicle dynamic energy is compared to the common dynamic level. Thus, when the vehicle dynamic energy changes drastically, the value of the vehicle's dynamic response adaptive adjustment factor increases accordingly. Furthermore, the dynamic effective threshold is proportionally increased through the dynamic response adaptive adjustment factor, thereby solving the problem that traditional fixed thresholds cannot adapt to vehicle dynamic changes and are prone to false alarms under high-dynamic working conditions. This allows for a larger normal fluctuation space for residual data when the vehicle is performing reasonable high-dynamic operations during fault detection. Finally, the third term in the formula for the dynamic response adaptive adjustment factor is an adaptive adjustment for the micro-fluctuation characteristics of the residual data of the wheel speed sensor. The standard deviation of the filtered residual data in the short term intuitively reflects the current noise level of the residual data. This kind of fluctuation may come from the electrical noise of the sensor itself, the minor irregularities during the induction with the toothed ring, or the subtle vibration transmission caused by road unevenness. By comparing the current real-time residual standard deviation with a reference residual fluctuation calibrated from healthy data to evaluate the cleanliness of the current data. If the volatility of the current residual data increases, this ratio will also increase, thereby enhancing the vehicle's dynamic response adaptive adjustment factor to increase the dynamic effective threshold. This enables the system to appropriately relax the judgment criteria when the quality of the residual data deteriorates or is affected by external minor interferences, avoiding misjudging normal noise fluctuations as faults, and thus enhancing the ability to preliminarily screen the vehicle wheel speed residual data.

[0037] Step S3: By performing a significance analysis on the consistency and amplitude of the wheel speed residual data under the low-dynamic interference state of the vehicle, a micro-drift confirmation factor of the vehicle is obtained.

[0038] After obtaining the vehicle's dynamic response adaptive adjustment factor in step S2, the set basic threshold of the vehicle wheel speed residual data can be dynamically adjusted through this vehicle's dynamic response adaptive adjustment factor. Specifically, the calculation result of multiplying the vehicle's dynamic response adaptive adjustment factor by the set basic threshold of the vehicle wheel speed residual data is used as the first dynamic threshold of the vehicle wheel speed residual data at the target moment. The obtained first dynamic threshold can effectively filter out most of the larger residuals caused by the normal dynamic behavior of the vehicle or significant signal noise. However, when there is a systematic deviation with a small amplitude and intermittent occurrence in the wheel speed sensor, the residual signal generated is often not large enough in amplitude to exceed the first dynamic threshold, resulting in the inability to identify the systematic deviation with a small amplitude and intermittent occurrence.

[0039] In traditional fault detection methods, whether it is a fixed threshold or only a dynamically adjusted threshold, the judgment basis mainly focuses on whether the instantaneous amplitude of the residual data exceeds the limit. Since the harm of such deviations is more reflected in their continuous deviation from a specific direction relative to normal random noise rather than a huge amplitude at a single moment, this method has limited detection ability for systematic deviations with small amplitudes but continuous existence, or small deviations that appear intermittently but have a clear bias when they appear.

[0040] Therefore, the present invention conducts preliminary screening of residual data through a first dynamic threshold, and uses the vehicle wheel speed residual data lower than the first dynamic threshold as candidate residual data, so as to obtain a micro-drift confirmation factor of the vehicle by performing a significance analysis on the consistency and amplitude of the candidate residuals under the vehicle's low-dynamic interference state.

[0041] First, evaluate the consistency of the candidate residual data to obtain a conditional cumulative consistency score; set the length of a long observation time window for judging micro-drift; divide the vehicle wheel speed residual data through the length of the long observation time window to obtain a first window of the vehicle wheel speed residual data; evaluate the driving state through the vehicle's real-time driving speed, lateral acceleration, and longitudinal acceleration to obtain an evaluation result of the vehicle's steady-speed driving condition. If the vehicle is in a steady-speed driving state, the evaluation result of the vehicle's steady-speed driving condition is 1. If the vehicle is not in a steady-speed driving state, the evaluation result of the vehicle's steady-speed driving condition is 0; use the candidate residual data that meets the evaluation result of the vehicle's steady-speed driving condition of 1 in the first window of the vehicle wheel speed residual data at the target moment as the second candidate wheel speed residual data at the target moment; use the absolute value of the sum of all the second candidate wheel speed residual data in the first window at the target moment as the cumulative deviation direction evaluation at the target moment; use the calculation result of the sum of the absolute values of all the second candidate wheel speed residual data in the first window at the target moment as the total fluctuation magnitude evaluation at the target moment; use the ratio of the cumulative deviation direction evaluation at the target moment to the total fluctuation magnitude evaluation at the target moment as the conditional cumulative consistency score of the candidate residual data at the target moment.

[0042] After that, analyze the amplitude change of the candidate residual data to obtain a relative degree score of the candidate residual data's conditional amplitude exceeding the baseline noise, and set a significance factor for the candidate residual; use the calculation result of multiplying the baseline noise standard deviation data by the significance factor as the significance standard deviation of the candidate residual; use the calculation result of subtracting the mean value of the absolute values of the second candidate wheel speed residual data at the target moment from the significance standard deviation and dividing by the baseline noise standard deviation data as the first relative degree evaluation at the target moment; use the larger value between the first relative degree evaluation at the target moment and the constant 0 as the relative degree score of the candidate residual data's conditional amplitude exceeding the baseline noise.

[0043] Finally, the calculation result of multiplying the conditional cumulative consistency score of the candidate residual data at the target moment by the relative degree score of the conditional amplitude of the candidate residual data exceeding the baseline noise is used as the micro-drift confirmation factor of the vehicle.

[0044] In one embodiment, assume that the candidate residual data at the -th moment is ; the long observation time window for judging micro-drift is ; the vehicle steady-speed driving condition indication function is ; in the long observation time window for judging micro-drift, when the vehicle is in the steady-speed driving condition, the average value of the absolute values of all candidate residual data is ; the baseline noise standard deviation data of the health state residual data of the vehicle under the steady-speed driving condition is , then the calculation formula for the micro-drift confirmation factor of the vehicle at the moment is:

[0045] Among them, represents the micro-drift confirmation factor of the vehicle at the moment; represents the long observation time window for judging micro-drift; represents the vehicle steady-speed driving condition indication function, which is used to judge the driving state of the vehicle. If the vehicle is in the predefined medium steady-speed driving state, then , otherwise ; represents the average value of the absolute values of all candidate residual data when the vehicle is in the steady-speed driving condition in the long observation time window for judging micro-drift; represents the significance factor. In the present invention, the significance factor is set to define how many times the average amplitude of the candidate residual needs to exceed the baseline noise standard deviation to be considered a statistically significant deviation; represents the baseline noise standard deviation data of the health state residual data of the vehicle under the steady-speed driving condition.

[0046] It should be noted that the core of the vehicle's micro-drift confirmation factor is to achieve precise identification of the micro-systematic deviation of the wheel speed sensor through a dual-criterion mechanism, thereby making up for the deficiencies of the traditional instantaneous amplitude judgment method. The first item is the conditional cumulative consistency score. The numerator represents the absolute value of the algebraic sum of all candidate residuals when the vehicle is in a specific constant-speed driving condition within a long observation window, reflecting the intensity of the net deviation direction accumulated by these residual data. The denominator represents the sum of the absolute values of all candidate residuals under the same condition, reflecting the total fluctuation magnitude of these residuals. The vehicle constant-speed driving condition indicator function is used to determine whether the vehicle is in a predefined medium constant-speed driving state at a historical moment. In this embodiment, when the real-time speed of the vehicle is between 30 km / h and 80 km / h, and both the longitudinal acceleration and the lateral acceleration of the vehicle are less than it is determined that the vehicle is in a medium constant-speed driving state. The value range of this conditional cumulative consistency score is between 0 and 1. When all candidate residual data that meet the conditions have the same sign, that is, when there is a continuous positive or negative deviation, this score approaches 1, indicating that the deviation has a high degree of one-way consistency; when the signs of the candidate residuals are randomly distributed, this score approaches 0. The main role of the conditional cumulative consistency score is to reveal at the statistical level whether candidate residuals show a continuous and single-direction deviation characteristic under specific driving conditions. This part aims to distinguish systematic small deviations with a fixed bias from background noise with random directions, providing a directional criterion for identifying potential fault signals. The second item in the vehicle's small drift confirmation factor is the relative degree score of the conditional amplitude exceeding the baseline noise, which further evaluates the actual abnormal degree of the small deviation with one-way consistency. This design compares the average amplitude of the candidate residuals with the pre-calibrated baseline noise data of the healthy state under this specific driving condition and introduces a significance factor to ensure that only when the deviation amplitude truly significantly exceeds the normal fluctuation range is it considered an abnormal with statistical significance, thereby reducing the misjudgment of normal small noise of the sensor and improving the accuracy of detection. The vehicle's small drift confirmation factor multiplies the evaluation results of the conditional cumulative consistency score and the relative degree score of the conditional amplitude exceeding the baseline noise to form a collaborative judgment logic. Only when the candidate residuals simultaneously meet the two conditions of continuous one-way deviation and significantly abnormal amplitude under specific driving conditions will the vehicle's small drift confirmation factor output a significant fault indication value.

[0047] Step S4: Perform continuous fault judgment by combining the vehicle's dynamic response adjustment factor and the small drift confirmation factor to obtain the fault determination result of the small systematic deviation of the wheel speed sensor.

[0048] After calculating the vehicle's dynamic response adaptive adjustment factor and obtaining the first dynamic threshold in step S2, and calculating the vehicle's small drift confirmation factor in step S3, in this step, based on the foregoing calculation results, a final determination is made on whether there is a small systematic deviation fault in the target sensor.

[0049] The first dynamic threshold obtained by adaptively adjusting the factor according to the vehicle's dynamic response is mainly used to preliminarily screen the residual data to adapt to the vehicle's real-time dynamic working conditions and define the candidate residual data that needs to be further analyzed by the vehicle's micro-drift confirmation factor. The core fault concerned in the present invention, that is, the micro-systematic deviation of the wheel speed sensor, its final identification and confirmation rely on the evaluation value of the vehicle's micro-systematic deviation. To ensure the accuracy of fault determination and avoid misjudgment caused by possible transient disturbances or single-point peaks in the output sequence of the vehicle's micro-drift confirmation factor, in this embodiment, when comparing the value of the vehicle's micro-drift confirmation factor with the threshold, a judgment logic of continuous confirmation is introduced. Specifically, a fault judgment threshold Threshold is set. In this embodiment, the fault judgment threshold Threshold is set to 0.7, and this threshold can be adjusted according to the actual scenario without requirements. A fault confirmation continuous cycle number M is set. In this embodiment, the value of the fault confirmation continuous cycle number M is set to 1000. Based on the system data sampling period, the value of the vehicle's micro-drift confirmation factor is calculated and updated in real time. A fault confirmation counter inside the vehicle is set. In each calculation cycle, if the value of the vehicle's micro-drift confirmation factor is greater than the fault judgment threshold, the fault confirmation counter inside the vehicle is incremented; if the value of the vehicle's micro-drift confirmation factor is not greater than the fault judgment threshold, the fault confirmation counter inside the vehicle is reset to 0. When the cumulative value of the fault confirmation counter reaches the continuous cycle number M, the system determines that the target wheel speed sensor has a micro-systematic deviation fault. Taking a calculation cycle of every 10 milliseconds as an example, M is equal to 1000, which means that the vehicle's micro-drift confirmation factor needs to be continuously higher than the fault judgment threshold for 10 seconds to confirm the fault.

[0050] Step S5, through signal reporting and status storage processing of the fault determination result, perform an active maintenance response for the wheel speed sensor.

[0051] After completing the determination of whether the target wheel speed sensor has a micro-systematic deviation fault in step S4, this step performs corresponding fault information processing and response according to the determination result.

[0052] If the determination result indicates that there is a minor systematic deviation fault in the target wheel speed sensor, the system will generate and record a diagnostic trouble code indicating such a specific early fault. This fault information will be transmitted through the vehicle communication network to the vehicle's central diagnostic management module, further triggering the corresponding warning indicator light on the dashboard to light up, warning the driver that there is a potential risk of performance degradation in the vehicle's electronic braking system. At the same time, the fault diagnostic code and relevant context information such as the fault occurrence time and vehicle operating status will be stored in the vehicle's non-volatile fault memory. These stored information provide an important basis for subsequent vehicle repair and maintenance, enabling maintenance technicians to accurately locate the potentially problematic wheel speed sensor and conduct targeted inspections, cleaning, calibration, tightening or timely replacement, thus achieving active intervention and predictive maintenance for the early performance degradation of the sensor, effectively preventing the further development and expansion of the fault.

[0053] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A real-time fault detection method for a vehicle electronic braking system, characterized in that, The real-time fault detection method for a vehicle electronic braking system includes the following steps: Step S1: Collect vehicle driving state data through an in-vehicle sensor network and an electronic control unit, calculate the filtered residual data of the target wheel speed sensor, and obtain the calibration reference parameters for the healthy state; Step S2: Through the joint analysis of vehicle dynamic energy and residual signal volatility on vehicle acceleration data and residual volatility data, obtain the dynamic response adaptive adjustment factor of the vehicle; Step S3: Through the significance analysis of the consistency and amplitude of wheel speed residual data under the low dynamic interference state of the vehicle, obtain the micro drift confirmation factor of the vehicle; Step S4: Through the combination of the dynamic response adjustment factor and the micro drift confirmation factor of the vehicle to conduct a continuous fault judgment, obtain the fault judgment result of the micro systematic deviation of the wheel speed sensor; Step S5: Through the signal reporting and status storage processing of the fault judgment result, conduct the active maintenance response of the wheel speed sensor.

2. The real-time fault detection method of a vehicle electronic braking system according to claim 1, characterized in that The collection of vehicle driving state data through an in-vehicle sensor network and an electronic control unit, the calculation of the filtered residual data of the target wheel speed sensor, and the establishment of the calibration reference parameters for the healthy state include: Set the data sampling period, collect vehicle driving state data through an in-vehicle sensor network and an electronic control unit, and the collected data includes: the longitudinal acceleration of the vehicle, the lateral acceleration of the vehicle, the vehicle reference speed, and the original wheel speed data output by the vehicle target wheel speed sensor; Obtain the expected wheel speed of the vehicle target wheel speed sensor through the vehicle reference speed and the dynamic tire radius of the target wheel; Take the difference between the original wheel speed data and the expected wheel speed as the original residual data of the vehicle target wheel speed; Filter the original residual data of the vehicle target wheel speed through a second-order Butterworth low-pass filter to obtain the residual data of the vehicle target wheel speed; Obtain the reference dynamic energy data of the normal dynamic energy level of the vehicle, the reference residual volatility data of the inherent volatility level of the residual signal when the wheel speed sensor is working normally, and the baseline noise standard deviation data of the residual data in the healthy state under the condition of steady-speed driving through the offline calibration process of the vehicle; Store the reference dynamic energy data, the reference residual volatility data, and the baseline noise standard deviation data as reference parameters in the electronic control unit of the vehicle.

3. The real-time fault detection method for a vehicle electronic braking system according to claim 1, characterized in that, The joint analysis of vehicle dynamic energy and residual signal volatility on vehicle acceleration data and residual volatility data to obtain the dynamic response adaptive adjustment factor of the vehicle includes: Through the vehicle dynamic energy analysis of the vehicle longitudinal acceleration and lateral acceleration, obtain the first vehicle dynamic energy index at the target moment; Set the time series window of the wheel speed sensor residual data, and obtain the short-term wheel speed residual standard deviation of the wheel speed sensor residual data in the previous time series window at the target moment; Through the fusion evaluation of the first vehicle dynamic energy index and the short-term wheel speed residual standard deviation on the dynamic driving intensity and the wheel speed residual volatility level, obtain the dynamic response adaptive adjustment factor of the vehicle.

4. The real-time fault detection method of a vehicle electronic braking system according to claim 3, characterized in that, Performing vehicle dynamic energy analysis on the longitudinal acceleration and lateral acceleration of the vehicle to obtain the first vehicle dynamic energy index at the target moment, including: Setting the length of the sliding time window for dynamic energy analysis; obtaining the sum of the squares of the lateral acceleration and longitudinal acceleration of the vehicle at each time stamp in the sliding time window for dynamic energy analysis; taking the square root of half of the mean of the sum of the squares of the acceleration corresponding to all time stamps in the sliding time window as the first vehicle dynamic energy index at the target moment.

5. The real-time fault detection method for a vehicle electronic braking system according to claim 3, characterized in that, Performing a fusion evaluation of the dynamic driving intensity and the wheel speed residual fluctuation level by using the first vehicle dynamic energy index and the short-term wheel speed residual standard deviation to obtain the dynamic response adaptive adjustment factor of the vehicle, including: Obtaining the reference dynamic energy data and the reference residual fluctuation data; taking the ratio of the first vehicle dynamic energy index to the reference dynamic energy data as the first adjustment factor; taking the ratio of the short-term wheel speed residual standard deviation to the reference residual fluctuation data as the second adjustment factor; taking the calculation result of adding the constant 1, the first adjustment factor and the second adjustment factor as the dynamic response adaptive adjustment factor of the vehicle.

6. The real-time fault detection method for a vehicle electronic braking system according to claim 1, characterized in that Performing a significance analysis on the consistency and amplitude of the wheel speed residual data under the low dynamic interference state of the vehicle to obtain the micro-drift confirmation factor of the vehicle, including: Setting the basic threshold of the vehicle wheel speed residual data, and performing weighted optimization on the basic threshold through the dynamic response adaptive adjustment factor of the vehicle to obtain the first dynamic threshold of the vehicle wheel speed residual data; performing preliminary screening on the vehicle wheel speed residual data through the first dynamic threshold of the vehicle wheel speed residual data to obtain the candidate residual data of the vehicle wheel speed; obtaining the conditional cumulative consistency score through the consistency evaluation of the candidate residual data; obtaining the relative degree score of the candidate residual data condition amplitude exceeding the baseline noise through the amplitude change analysis of the candidate residual data; obtaining the micro-drift confirmation factor of the vehicle through the comprehensive evaluation of the conditional cumulative consistency score and the relative degree score of the candidate residual data condition amplitude exceeding the baseline noise.

7. The real-time fault detection method for a vehicle electronic braking system according to claim 6, characterized in that, Obtaining the conditional cumulative consistency score of the candidate residual data through the consistency evaluation of the candidate residual data; obtaining the relative degree score of the candidate residual data condition amplitude exceeding the baseline noise through the amplitude change analysis of the candidate residual data, including: Setting the length of the long observation time window for judging micro-drift; performing window division on the vehicle wheel speed residual data through the length of the long observation time window to obtain the first window of the vehicle wheel speed residual data; obtaining the evaluation result of the vehicle steady speed driving condition through the real-time driving speed, lateral acceleration and longitudinal acceleration of the vehicle. If the vehicle is in the steady speed driving state, the evaluation result of the vehicle steady speed driving condition is 1. If the vehicle is not in the steady speed driving state, the evaluation result of the vehicle steady speed driving condition is 0; The candidate residual data that meets the evaluation result of 1 for the vehicle steady-speed driving condition in the first window of the vehicle wheel speed residual data at the target moment is used as the second candidate wheel speed residual data at the target moment; the absolute value of the sum of all the second candidate wheel speed residual data in the first window at the target moment is used as the cumulative deviation direction evaluation at the target moment; the calculation result of the sum of the absolute values of all the second candidate wheel speed residual data in the first window at the target moment is used as the total fluctuation magnitude evaluation at the target moment; the ratio of the cumulative deviation direction evaluation at the target moment to the total fluctuation magnitude evaluation at the target moment is used as the conditional cumulative consistency score of the candidate residual data at the target moment. Set the significance factor of the candidate residual; the calculation result of multiplying the baseline noise standard deviation data by the significance factor is used as the significance standard deviation of the candidate residual; the calculation result of subtracting the mean value of the absolute values of the second candidate wheel speed residual data at the target moment from the significance standard deviation and then dividing by the baseline noise standard deviation data is used as the first relative degree evaluation at the target moment; the larger value between the first relative degree evaluation at the target moment and the constant 0 is used as the relative degree score of the candidate residual data condition amplitude exceeding the baseline noise.

8. The real-time fault detection method of a vehicle electronic braking system according to claim 6, characterized in that, The comprehensive evaluation through the conditional cumulative consistency score and the relative degree score of the vehicle wheel speed residual data condition amplitude exceeding the baseline noise to obtain the micro drift confirmation factor of the vehicle includes: The calculation result of multiplying the conditional cumulative consistency score of the candidate residual data at the target moment by the relative degree score of the candidate residual data condition amplitude exceeding the baseline noise is used as the micro drift confirmation factor of the vehicle.

9. A real-time fault detection method for a vehicle electronic braking system according to claim 1, characterized in that, The continuous fault judgment by combining the dynamic response adjustment factor of the vehicle and the micro drift confirmation factor to obtain the fault judgment result of the micro systematic deviation of the wheel speed sensor includes: Set the fault judgment threshold and the fault confirmation duration length, initialize the fault confirmation counter. In the data sampling period, if the micro drift confirmation factor of the vehicle is greater than the fault judgment threshold, the fault confirmation counter is incremented; if the micro drift confirmation factor of the vehicle is less than or equal to the fault judgment threshold, the fault confirmation counter is reset to 0; when the cumulative value of the fault confirmation counter reaches the duration length, it is determined that the target wheel speed sensor has a micro systematic deviation fault.

Images