A brake pad spring crack data analysis method and system

Abstract

The application provides a brake pad spring crack data analysis method and system, relates to the field of vehicle fault diagnosis and data analysis, receives third-party maintenance station maintenance report, vehicle-mounted vibration sensor data and user feedback data; identifies contradictions among the three and vehicle engine stall, issues standardized brake instructions to drive brake caliper clamping, so that brake pad assembly containing brake pad spring is excited to generate vibration response; synchronously triggers an accelerometer to collect target vibration response signals at a high sampling rate; performs advanced feature extraction on the signals to obtain a vehicle feature vector; adopts a multi-dimensional feature vector similarity comparison algorithm to compare the feature vector with a preset standard sample library feature set, and determines the real crack state of the brake pad spring according to the comparison result. The application solves the brake pad spring crack misjudgment problem under the traditional single data source by triggering standardized excitation through multi-source data contradictions and combining high-fidelity vibration signal feature analysis, and improves the detection accuracy under complex working conditions.

Description

Technical Field

[0001] This application relates to the field of vehicle fault diagnosis and data analysis, and more specifically, to a method and system for analyzing brake pad spring crack data. Background Technology

[0002] As a critical safety component of a vehicle's braking system, the integrity of brake pad springs directly impacts driving safety. Currently, maintenance services for shared vehicles typically rely on numerous third-party repair shops. These repair shops vary significantly in size, technical expertise, and equipment configuration. For example, large chain repair centers may be equipped with advanced diagnostic equipment capable of providing high-precision, structured data, such as detection systems based on ultrasonic arrays or eddy current imaging principles, which can accurately measure the length, width, and depth of cracks. In contrast, small, independent repair shops, limited by cost, generally still rely on traditional magnetic particle testing equipment. While magnetic particle testing technology is less expensive, its ability to detect microcracks is limited, and the results are mostly qualitative, such as "whether there is a crack indication." This difference in equipment and technical expertise leads to significant inconsistencies in the accuracy, format, and information dimensions of brake pad spring crack detection data uploaded by different repair shops to the central database of the car-sharing platform.

[0003] Furthermore, human error exacerbates data inaccuracies. For example, when repair shops face stringent vehicle turnover targets, technicians may omit crucial equipment calibration steps, such as the daily mandatory calibration of magnetic particle testing equipment, to save time. This negligence directly leads to a significant decrease in the sensitivity of the testing equipment, making it impossible to effectively detect microscopic metal fatigue cracks that are difficult to detect with the naked eye. When such potentially hazardous brake pads and springs are incorrectly marked as "inspection qualified" and uploaded to the platform database, erroneous "normal" status data is introduced.

[0004] These erroneous maintenance data severely impacted the crack condition assessment system of the car-sharing platform. This system was designed to predict the health of brake pad springs by analyzing maintenance data and onboard sensor data. However, when the system received real crack samples incorrectly labeled as "normal," its internal assessment logic and parameters were inaccurately adjusted. This caused the system to misjudge vehicles with similar characteristics in subsequent risk assessments, mistakenly assuming their braking systems had low risk.

[0005] When the vehicle is put back into operation, although the micro-cracks in the brake pad springs continue to expand under frequent braking and complex road conditions, changing their structural integrity and vibration patterns, the onboard vibration sensors begin to collect abnormal vibration waveforms with specific frequency characteristics. However, because the platform's backend analysis system has been affected by erroneous information, its evaluation logic cannot accurately identify the actual crack information represented by these abnormal vibration waveforms. The system may incorrectly identify these real crack warning signals as "instantaneous impact from speed bumps" or "normal vibration caused by road bumps," marking them as benign events requiring no human intervention, thus ignoring potential safety hazards.

[0006] Furthermore, user feedback, as an important source of information, presents new challenges due to its unstructured and subjective nature. When users perceive signs of malfunction such as "soft brakes" or "strange noises" and report them through the platform, these natural language descriptions differ significantly from structured sensor data and maintenance records. At this point, the platform's vehicle dispatch system faces conflicting information from different sources: structured maintenance data from repair stations shows "normal," vehicle vibration sensor data is misinterpreted as "normal," while user feedback clearly points to a malfunction. Lacking a mechanism to effectively integrate, calibrate, and interpret these heterogeneous and contradictory information sources, the system cannot issue recall repair orders based on any single information source or simple logical rules. This forces vehicles with safety hazards to continue operating online without timely and effective intervention. Summary of the Invention

[0007] This application provides a method and system for analyzing brake pad spring crack data. Through active excitation, high-precision data acquisition, and multi-source data fusion analysis, it effectively solves the problem of misjudgment of brake pad spring cracks caused by inconsistent and inaccurate data in the prior art, thereby improving the accuracy and reliability of diagnosis.

[0008] On the one hand, this application provides a method for analyzing brake pad spring crack data, including:

[0009] Receive maintenance reports from third-party repair shops, vehicle vibration sensor data, and user feedback data;

[0010] When a discrepancy is detected between the maintenance report, the vehicle vibration sensor data, and the user feedback data, and the vehicle engine is off, a standardized braking command is issued to the target vehicle's braking system to drive the brake caliper to perform a clamping action, causing the brake pad spring to receive standardized excitation and generate a vibration response; wherein, the brake caliper clamps the brake pad assembly, and the brake pad assembly includes the brake pad spring and the brake friction pad;

[0011] Simultaneously with the issuance of the standardized braking command, a synchronous trigger signal is sent to the accelerometer, which acquires the target vibration response signal generated by the brake pad spring under the standardized excitation at a preset high sampling rate;

[0012] Advanced feature extraction is performed on the target vibration response signal to obtain vehicle feature vector information;

[0013] A multidimensional feature vector similarity comparison algorithm is used to compare the vehicle feature vector information with the feature set in a preset standard sample library to obtain the comparison result. Based on the comparison result, the true crack state of the brake pad spring is determined.

[0014] Optionally, before issuing a standardized braking command to the braking system of the target vehicle to drive the brake caliper to perform a clamping action, the method further includes:

[0015] A standardized physical surface pretreatment command is issued to the braking system, driving the brake caliper to perform multiple clamping and releasing operations on the brake pad spring according to the set pressure and duration; wherein, the clamping and releasing operations cause the brake friction pads in the brake pad assembly to rub against the brake disc, so as to remove the unstable oxide film on the friction surface.

[0016] Optionally, the step of performing high-level feature extraction on the target vibration response signal to obtain vehicle feature vector information includes:

[0017] The target vibration response signal is subjected to three-level wavelet packet decomposition to obtain sub-frequency band signals;

[0018] Extract the time-domain and frequency-domain feature parameters of each sub-band signal, wherein: the time-domain feature parameters include: root mean square value, kurtosis coefficient, and impulse factor; the frequency-domain feature parameters include: main band energy proportion and band energy entropy;

[0019] The time-domain feature parameters and frequency-domain feature parameters are combined into a multi-dimensional feature vector, which serves as the vehicle feature vector information.

[0020] Optionally, the step of sending a synchronous trigger signal to the accelerometer simultaneously with the issuance of the standardized braking command, and the accelerometer acquiring the target vibration response signal generated by the brake pad spring under the standardized excitation at a preset high sampling rate, includes:

[0021] Repeat the standardized stimulus a preset number of times;

[0022] Each time the standardized excitation is applied, the vibration response signal generated by the brake pad spring under the standardized excitation is collected synchronously to obtain multiple vibration response signals;

[0023] Time-align multiple vibration response signals and perform superposition to suppress noise, obtaining a target vibration response signal.

[0024] Optionally, the feature sets in the preset standard sample library include a standard healthy brake leaf spring feature set and a standard cracked brake leaf spring feature set; the step of comparing the vehicle feature vector information with the feature sets in the preset standard sample library using a multi-dimensional feature vector similarity comparison algorithm to obtain a comparison result, and determining the true crack state of the brake leaf spring according to the comparison result specifically includes:

[0025] Calculate the average distance DH between the vehicle feature vector information and all vectors in the standard healthy brake leaf spring feature set, and the average distance DC between the vehicle feature vector information and all vectors in the standard cracked brake leaf spring feature set;

[0026] If DC < DH and DC is less than a preset threshold, it is determined that the brake leaf spring is in a cracked state;

[0027] If DH < DC and DH is less than a preset threshold, it is determined that the brake leaf spring is in a healthy state.

[0028] Optionally, after the step of receiving the maintenance report from a third-party repair station, the on-vehicle vibration sensor data, and the user feedback data, it includes:

[0029] Extract the status from the maintenance report, the on-vehicle vibration sensor data, and the user feedback data, obtaining a status indication corresponding to each data source;

[0030] Assign a reliability weight to the status indication according to the type and historical performance of each data source;

[0031] Calculate the tendency score of each data source for the state of the brake leaf spring according to the status indication and the reliability weight;

[0032] Compare the tendency scores, and when there is a preset difference between the tendency scores, identify the contradiction existing among the maintenance report, the on-vehicle vibration sensor data, and the user feedback data.

[0033] Optionally, the step of assigning a reliability weight to the status indication according to the type and historical performance of each data source includes:

[0034] Assign an initial reliability weight to the status indication according to the type and historical performance of each data source;

[0035] Continuously monitor the consistency between the status indication of each data source and the true crack state of the brake leaf spring;

[0036] Based on the consistency, the initial reliability weight is adjusted to obtain the adjusted reliability weight, and the adjusted reliability weight is used as the reliability weight of the state indication.

[0037] Optionally, the step of extracting the status from the maintenance report, the vehicle vibration sensor data, and the user feedback data to obtain the status indication corresponding to each data source includes:

[0038] The maintenance report is subjected to structured information parsing and keyword extraction, and the parsing results are mapped to the status indication of the brake pad spring;

[0039] The time-domain or frequency-domain features of the vehicle vibration sensor data are extracted, and the extracted features are compared with a preset feature pattern to obtain the state indication of the brake pad spring.

[0040] In addition, natural language processing is performed on the user feedback data to identify keywords or phrases related to the state of the brake pad spring, and the identification results are mapped to the state indication of the brake pad spring.

[0041] Optionally, the step of calculating the tendency score of each data source for the brake spring state includes:

[0042] The status indicators of each data source are converted into numerical values, and the numerical values ​​are weighted and summed or weighted averaged according to the reliability weights to obtain the propensity score.

[0043] On the other hand, this application provides a brake pad spring crack data analysis system, the system comprising:

[0044] The data receiving module is used to receive maintenance reports from third-party repair shops, vehicle vibration sensor data, and user feedback data.

[0045] An excitation control module is used to issue a standardized braking command to the braking system of the target vehicle when a contradiction is detected between the maintenance report, the vehicle vibration sensor data, and the user feedback data, and the vehicle engine is off. This command drives the brake caliper to perform a clamping action, causing the brake pad spring to receive standardized excitation and generate a vibration response. The brake caliper clamps the brake pad assembly, which includes a brake pad spring and brake friction pads.

[0046] The vibration response acquisition module is used to send a synchronous trigger signal to the accelerometer at the same time as the standardized braking command is issued. The accelerometer acquires the target vibration response signal generated by the brake pad spring under the standardized excitation at a preset high sampling rate.

[0047] Feature extraction module: performs advanced feature extraction on the target vibration response signal to obtain vehicle feature vector information;

[0048] The crack diagnosis module uses a multi-dimensional feature vector similarity comparison algorithm to compare the vehicle feature vector information with the feature sets in the preset crack sample library and health sample library. When the similarity with the crack sample library exceeds a first threshold, a crack diagnosis conclusion is output; when the similarity with the health sample library exceeds a second threshold, a health diagnosis conclusion is output.

[0049] This application provides a method and system for analyzing brake pad spring crack data. Through active excitation, high-precision data acquisition, and multi-source data fusion analysis, it effectively solves the problem of misjudgment of brake pad spring cracks caused by inconsistent and inaccurate data in the prior art, thereby improving the accuracy and reliability of diagnosis. Attached Figure Description

[0050] To illustrate this application more clearly, the accompanying drawings used in the embodiments will be briefly described below. Obviously, those skilled in the art can obtain other drawings based on these drawings without any creative effort.

[0051] Figure 1 The diagram above illustrates a flowchart of a brake pad spring crack data analysis method in an embodiment.

[0052] Figure 2 The diagram illustrates a module configuration block diagram of a brake pad spring crack data analysis system in an embodiment.

[0053] Figure reference numerals: 100, Brake pad spring crack data analysis system; 10, Data receiving module; 20, Excitation control module; 30, Vibration response acquisition module; 40, Feature extraction module; 50, Crack diagnosis module. Detailed Implementation

[0054] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0055] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0056] Traditional shared vehicle brake pad spring condition assessment systems suffer from heterogeneous data sources, conflicting information, and a lack of a single, reliable source when integrating maintenance reports from third-party repair shops, onboard vibration sensor data, and user feedback. Specifically, differences in testing equipment and operating procedures among different repair shops lead to inconsistencies in the accuracy and format of maintenance data. For example, small repair shops using magnetic particle testing equipment may fail to detect microscopic cracks, while large repair shops using high-precision image scanning equipment can provide detailed, structured data. Furthermore, onboard vibration sensor data may misinterpret data due to contamination of the assessment model with erroneous maintenance data, misidentifying real crack signals as benign events. Additionally, user feedback data is typically unstructured textual descriptions, making standardization and quantitative analysis difficult. These problems prevent the system from accurately identifying and confirming the true crack condition of the brake pad spring, thus impacting decisions regarding vehicle safety operations.

[0057] If the aforementioned issues are not addressed, car-sharing platforms will continue to face the risk of misjudging the condition of brake pad spring cracks. Such misjudgments could lead to vehicles with safety hazards continuing to operate online, threatening user safety. Furthermore, the failure to promptly detect and repair cracks could further exacerbate the damage to the brake pad springs, ultimately causing component failure and triggering more serious traffic accidents. In addition, erroneous assessment results will continuously contaminate the platform's risk assessment model, causing a chain reaction of misjudgments in subsequent vehicle condition assessments, reducing the reliability and accuracy of the entire system. In the long run, this will not only damage user trust and brand reputation but also increase operating costs, such as compensation for accidents, vehicle repair expenses, and potential legal liabilities.

[0058] like Figure 1 The diagram illustrates an exemplary method for analyzing brake pad spring crack data. By integrating data from multiple sources and applying standardized excitation, a reliable vibration response is obtained, thereby analyzing the crack state. This application proposes a brake pad spring crack data analysis method, comprising:

[0059] The S10 receives maintenance reports from third-party repair shops, data from onboard vibration sensors, and user feedback data.

[0060] S20, when a contradiction is detected between the maintenance report, the vehicle vibration sensor data, and the user feedback data, and the vehicle engine is off, a standardized braking command is issued to the target vehicle's braking system to drive the brake caliper to perform a clamping action, so that the brake pad spring receives standardized excitation to generate a vibration response; wherein, the brake caliper clamps the brake pad assembly, and the brake pad assembly includes a brake pad spring and a brake friction pad.

[0061] Standardized braking commands refer to a preset, repeatable braking operation command. It can be implemented by the vehicle control unit (ECU) issuing specific braking pressure commands through bus communication, or by simulating brake pedal signals through external testing equipment. It can ensure that the brake pad springs receive consistent excitation under controlled conditions, thereby eliminating the influence of environmental factors or operational differences on vibration response data.

[0062] Under the drive of standardized braking commands, the brake pad spring is subjected to a predetermined pattern of force or displacement, thereby generating predictable vibrations. This can be achieved by using brake calipers to precisely clamp and release the brake pad assembly, simulating the stress state of the brake pad spring under actual braking conditions and inducing it to generate specific vibration modes when cracks are present.

[0063] A brake pad assembly, comprising a brake pad spring and brake friction pads, refers to the component in a braking system that directly contacts the brake disc and generates frictional force; it is a physical object used to define a standardized excitation effect.

[0064] S30, at the same time as the standardized braking command is issued, a synchronous trigger signal is sent to the accelerometer, and the accelerometer acquires the target vibration response signal generated by the brake pad spring under the standardized excitation at a preset high sampling rate.

[0065] The synchronous trigger signal is a time synchronization signal sent to the accelerometer at the same time as the standardized braking command is issued. It can be implemented by hardware trigger line connection or software timestamp synchronization to ensure that the accelerometer starts data acquisition at the precise moment of excitation application and ensures that the vibration response signal is aligned with the excitation action.

[0066] S40, perform advanced feature extraction on the target vibration response signal to obtain vehicle feature vector information.

[0067] Complex mathematical transformations and processing are performed on the original vibration response signal to achieve advanced feature extraction, which extracts feature parameters that can characterize the state of the brake spring with lower dimensionality but richer information. Various signal processing techniques such as time domain analysis, frequency domain analysis, and time-frequency analysis can be used to transform the original and massive vibration data into structured information that can be used for pattern recognition and comparison.

[0068] Vehicle feature vector information refers to a multi-dimensional set of values ​​that represents the vibration characteristics of the brake pad spring of the current vehicle after advanced feature extraction, and is used as input for subsequent similarity comparison.

[0069] S50, a multi-dimensional feature vector similarity comparison algorithm is used to compare the vehicle feature vector information with the feature set in the preset standard sample library to obtain the comparison result, and the actual crack state of the brake pad spring is determined based on the comparison result.

[0070] Multidimensional feature vector similarity comparison algorithm refers to a calculation method used to quantify the similarity between two or more multidimensional feature vectors. It can be implemented using various distance or similarity measurement methods such as Euclidean distance, cosine similarity, and Mahalanobis distance. It can objectively evaluate the degree of similarity between the features of the brake pad spring under test and the features of the known healthy or cracked state.

[0071] The core innovation of this application lies in introducing a controlled, standardized physical excitation test when multi-source data contradictions are identified, and combining it with high-precision vibration response acquisition and advanced feature comparison, thereby overcoming the unreliability and contradictions of traditional data sources, achieving accurate judgment of the true crack state of brake pad springs, and achieving the effect of improving diagnostic reliability and preventing vehicles with safety hazards from continuing to operate.

[0072] This application establishes a multi-dimensional data input foundation by receiving maintenance reports from third-party repair shops, vehicle vibration sensor data, and user feedback data. When the system detects inconsistencies among these heterogeneous data sources—for example, a maintenance report showing normal data while vehicle data or user feedback indicates anomalies, and the vehicle is in a quiet state with the engine off—the system proactively issues a standardized braking command to the target vehicle's braking system. This command drives the brake calipers to perform precise clamping actions, ensuring the brake pad springs receive standardized excitation. This controlled excitation method eliminates the influence of uncertainties such as vehicle operating status, road conditions, or human operation on the vibration response, ensuring consistent conditions for each test. Simultaneously with the issuance of the standardized braking command, the system sends a synchronous trigger signal to the accelerometer, ensuring that the accelerometer begins acquiring the target vibration response signal generated by the brake pad springs at a preset high sampling rate at the precise moment the excitation is applied. The high sampling rate guarantees the ability to capture vibration details, providing high-quality raw data for subsequent analysis. Subsequently, advanced feature extraction is performed on the acquired target vibration response signal, transforming it into vehicle feature vector information. This process compresses the complex raw signal into a more representative set of values, reducing data dimensionality while preserving key vibration characteristics. Finally, a multi-dimensional feature vector similarity comparison algorithm is used to compare the extracted vehicle feature vector information with a feature set in a pre-defined standard sample library. The standard sample library contains brake spring features with known health and known crack states. Through the comparison results, the system can objectively determine the true crack state of the current brake spring, thus overcoming the limitations of a single data source and the interference of contradictory information.

[0073] For example, data reception can be achieved through a vehicle remote diagnostic interface or a cloud platform API. Maintenance reports, on-board vibration sensor data, and user feedback data are aggregated into a central processing unit (CPU). This CPU runs a data consistency check module that uses preset logical rules or machine learning models to evaluate whether there are significant discrepancies in the brake pad spring status indications from different data sources. Once a discrepancy is identified, and the vehicle's engine status sensor confirms that the engine is off, the CPU sends a standardized braking command to the brake control unit via the vehicle communication network. This command can be set to cause the brake caliper to perform one or more clamping actions with a specific pressure and duration. Simultaneously, a hardware trigger or software synchronization mechanism sends a synchronization signal to an accelerometer mounted near the brake caliper or brake pad assembly. This accelerometer can be a microelectromechanical system (MEMS) sensor that continuously acquires vibration data at a sampling rate of tens of thousands of times per second. The acquired raw vibration signals are transmitted to a signal processing module, which performs advanced feature extraction, such as extracting the dominant frequency band energy through frequency domain analysis or calculating the root mean square value through time domain analysis, combining these parameters into a multi-dimensional feature vector. Subsequently, the feature vector is fed into the pattern recognition module, which stores pre-trained standard healthy sample feature sets and standard crack sample feature sets. The pattern recognition module uses a similarity comparison algorithm to calculate the distance or similarity between the feature vector of the vehicle under test and each feature set in the standard sample library, and outputs a diagnosis of the health or crack status of the brake pad spring based on the comparison results.

[0074] Through the above technical solution, this application can effectively integrate and calibrate heterogeneous and contradictory information from different sources. By introducing controlled and standardized physical excitation, the inherent defects and uncertainties of traditional data sources are overcome, ensuring the reliability and consistency of vibration response data. Furthermore, by performing advanced feature extraction on high-precision vibration response signals and comparing their similarity with a standard sample library, accurate judgment of the true crack state of brake pad springs is achieved. This significantly improves the reliability and accuracy of brake pad spring crack diagnosis, preventing vehicles with potential safety hazards from continuing to operate, thereby ensuring vehicle operation safety.

[0075] In some embodiments, prior to the step of issuing a standardized braking command to the braking system of the target vehicle to drive the brake caliper to perform a clamping action, the method further includes:

[0076] A standardized physical surface pretreatment command is issued to the braking system, driving the brake caliper to perform multiple clamping and releasing operations on the brake pad spring according to the set pressure and duration; wherein, the clamping and releasing operations cause the brake friction pads in the brake pad assembly to rub against the brake disc, so as to remove the unstable oxide film on the friction surface.

[0077] Among these measures, cleaning and preparing the brake friction surfaces before formal braking excitation can be achieved using pre-programmed electronic control unit (ECU) instructions or specific commands issued through the on-board diagnostic (OBD) interface, ensuring the stability and repeatability of subsequent braking operations.

[0078] This application adds a physical surface pretreatment step before the step of issuing standardized braking commands to the braking system of the target vehicle to drive the brake caliper to perform clamping action. Specifically, before the formal standardized braking excitation, the system issues a standardized physical surface pretreatment command to the braking system, driving the brake caliper to perform multiple clamping and releasing operations on the brake pad spring according to a pre-set pressure and duration. This series of clamping and releasing operations causes repeated friction between the brake friction pads and the brake disc in the brake pad assembly. This friction can effectively remove unstable oxide films or other contaminants that may exist on the friction surface. The presence of these unstable substances will cause the clamping force and friction coefficient of the brake caliper to be unstable when performing the clamping action, resulting in deviations in the vibration response signal generated by the brake pad spring under standardized excitation, affecting the accuracy of subsequent crack state analysis.

[0079] By preemptively eliminating these unstable factors, a clean and stable baseline for the contact state between the brake pads and the brake disc can be ensured. Based on this, when subsequent standardized braking commands are issued, the clamping action of the brake caliper will be more stable and consistent, and the excitation on the brake pad spring will be more standardized, resulting in a purer and more reliable vibration response signal. This preprocessing mechanism, combined with subsequent standardized excitation acquisition and data analysis steps, forms a more robust brake pad spring crack diagnosis process.

[0080] By eliminating the uncertainty of the initial friction surface, the signal-to-noise ratio and repeatability of the acquired vibration response signal can be significantly improved. This, in turn, enhances the accuracy of advanced feature extraction of the target vibration response signal and the determination of the true crack state of the brake spring using a multi-dimensional feature vector similarity comparison algorithm. This avoids misjudgment caused by the instability of the friction surface and ensures the reliability of the diagnostic results.

[0081] For example, physical surface pretreatment can be performed before issuing standardized braking commands to the braking system of the target vehicle. Specifically, the system can send a pretreatment command to the vehicle's Electronic Brake Control Unit (EBCU), which includes a preset braking pressure value, such as 500 kPa, the duration of each clamping operation, such as 2 seconds, and the number of repetitions, such as 5 times. Upon receiving the command, the EBCU controls the brake hydraulic system to drive the brake calipers to clamp the brake pad assembly at a pressure of 500 kPa, hold for 2 seconds, and then release, repeating this process 5 times. During each clamping process, the brake pads rub against the brake disc, which can effectively wear away or peel off trace amounts of rust, oil, or old oxide layers adhering to the friction surface. For example, if the vehicle has been parked for a long time, a thin layer of surface rust may form on the brake disc surface; this can be removed through repeated light braking operations, restoring the friction surface to a relatively clean and stable state. After pretreatment, the system then issues the formal standardized braking command for subsequent vibration response signal acquisition and crack analysis.

[0082] By employing the aforementioned technical solution, standardized physical surface pretreatment can effectively remove unstable oxide films or other contaminants from the friction surfaces of the brake pads and brake discs before analyzing brake pad spring crack data. This results in more stable and consistent clamping action of the subsequent brake caliper, and the brake pad spring can generate a purer and more repeatable vibration response signal when subjected to standardized excitation. Therefore, it can eliminate the interference of unstable factors on the friction surface on the vibration response signal, improve the standardization of the braking process, and thus enhance the accuracy of brake pad spring crack state analysis, avoiding misjudgments.

[0083] In some embodiments, the step of performing high-level feature extraction on the target vibration response signal to obtain vehicle feature vector information includes:

[0084] The target vibration response signal is subjected to three-level wavelet packet decomposition to obtain sub-frequency band signals;

[0085] Extract the time-domain and frequency-domain feature parameters of each sub-band signal, wherein: the time-domain feature parameters include: root mean square value, kurtosis coefficient, and impulse factor; the frequency-domain feature parameters include: main band energy proportion and band energy entropy;

[0086] The time-domain feature parameters and frequency-domain feature parameters are combined into a multi-dimensional feature vector, which serves as the vehicle feature vector information.

[0087] The original signal is decomposed at different frequency scales to obtain a series of sub-signals with different frequency ranges, achieving three-level wavelet packet decomposition. Here, "three-level" refers to the number of decomposition levels; the higher the level, the more finely the signal is decomposed, allowing for more precise capture of local signal features. This effectively separates complex information from the original vibration response signal, facilitating subsequent feature extraction for specific frequency ranges and reducing noise interference during feature extraction.

[0088] Sub-band signals are obtained by wavelet packet decomposition and represent the components of the original signal within a specific frequency range. Each sub-band signal contains the energy and waveform information of the original signal within the corresponding frequency range. This is used to distribute the energy and information of the original signal to different frequency channels, so that the characteristics of different frequency components can be analyzed in a targeted manner, thereby more effectively identifying weak signals related to brake pad spring cracks.

[0089] Temporal characteristic parameters are numerical values ​​extracted from the signal's behavior on the time axis. They are used to quantify the signal's instantaneous characteristics, energy intensity, or waveform morphology. Specifically, the root mean square (RMS) value measures the effective energy or intensity of the signal, reflecting its overall vibration level. The kurtosis coefficient describes the degree to which the signal waveform deviates from a normal distribution; it is sensitive to impulsive or abrupt signals and can reflect the presence of abnormal impulsive components in the signal. The impulse factor reflects the relative intensity of the impulsive component in the signal and is useful for identifying periodic impacts or transient fault signals. The purpose of extracting these parameters is to capture the energy changes, impact characteristics, and waveform sharpness of the brake spring vibration response signal in the time dimension. These characteristics are often closely related to the initiation and propagation of cracks.

[0090] Frequency domain characteristic parameters refer to the values ​​extracted from the signal's performance on the frequency axis, used to quantify the signal's frequency distribution, energy concentration, or spectral complexity. Specifically, the main frequency band energy ratio measures the degree of concentration of signal energy in the main frequency range, and can reflect the main vibration mode of the signal.

[0091] Frequency band energy entropy is used to describe the uniformity or complexity of the energy distribution of a signal across different frequency components. The more uneven the energy distribution, the lower the entropy value, and vice versa. The purpose of extracting these parameters is to analyze the energy distribution and complexity of the brake spring vibration response signal from the frequency dimension, because the presence of cracks will change the natural frequency and vibration mode of the component, thus exhibiting specific changes in the frequency domain.

[0092] A multidimensional feature vector is a numerical sequence formed by combining multiple different types of feature parameters. It is used to integrate signal features extracted from different dimensions to form a comprehensive and discriminative mathematical representation, so that subsequent pattern recognition and classification algorithms can process it and thus more accurately represent the true state of the brake spring.

[0093] This application effectively solves the problem of inaccurate feature extraction caused by noise and redundant information in the original signal by performing multi-level and multi-dimensional processing on the target vibration response signal.

[0094] Specifically, firstly, the target vibration response signal is subjected to three-level wavelet packet decomposition, thereby decomposing the original signal into multiple sub-frequency band signals. This decomposition enables detailed analysis of the signal at different frequency scales, effectively separating key information related to the brake spring crack, while suppressing noise and interference components that may exist in the signal, providing a cleaner and more focused data foundation for subsequent feature extraction.

[0095] Based on this, time-domain and frequency-domain feature parameters are extracted from these sub-band signals. Time-domain feature parameters capture the signal's energy, impact characteristics, and waveform morphology over time, which are crucial for identifying transient or nonlinear vibration modes caused by cracks. Simultaneously, frequency-domain feature parameters reveal the signal's energy distribution and complexity over frequency, as cracks in the brake spring alter its natural frequency and vibration response spectrum. By extracting both types of features simultaneously, a more comprehensive and in-depth understanding of the vibration signal can be obtained, thereby enhancing the discriminative power of the feature vectors.

[0096] Finally, these time-domain and frequency-domain feature parameters are combined into a multi-dimensional feature vector, which serves as the vehicle's feature vector information. This combination fully utilizes the complementarity between different types of features, enabling the final feature vector to more robustly and accurately characterize the true vibration characteristics of the brake pad spring.

[0097] This advanced feature extraction method makes the vehicle feature vector information obtained from the target vibration response signal more representative and discriminative. When these optimized vehicle feature vectors are compared with the feature set in a pre-set standard sample library, the accuracy of the comparison results can be significantly improved, thus more accurately determining the true crack state of the brake pad spring. This method avoids the inaccuracies that may arise from directly analyzing the original signal, providing reliable data support for the crack diagnosis of brake pad springs, and effectively solving the problem of inaccurate feature extraction affecting the accuracy of crack state judgment.

[0098] For example, advanced feature extraction of a target vibration response signal can be implemented as follows: First, the target vibration response signal can be input into a signal processing unit, which can be an embedded processor or an on-board diagnostic computer. In this signal processing unit, a wavelet packet transform algorithm can be used to perform a three-level decomposition of the signal. For example, Daubechies wavelet basis functions or Symlets wavelet basis functions can be selected to perform a three-level decomposition of the original vibration signal, thereby generating eight different sub-frequency band signals, each corresponding to a specific frequency range. Subsequently, for each generated sub-frequency band signal, its time-domain feature parameters and frequency-domain feature parameters are calculated. The calculation of the time-domain feature parameters can include: the root mean square value, which can be obtained by averaging the squared values ​​of the signal and then taking the square root; the kurtosis coefficient, which can be obtained by calculating the ratio of the fourth central moment of the signal to the square of the second central moment; and the impulse factor, which can be obtained by the ratio of the peak value of the signal to the root mean square value. The calculation of frequency domain feature parameters can include: the energy proportion of the main frequency band, which can be obtained by identifying the frequency band with the most concentrated energy in the signal spectrum and calculating the proportion of that frequency band's energy to the total energy; and the frequency band energy entropy, which can be obtained by calculating the Shannon entropy of the probability distribution of the energy in each sub-frequency band to quantify the uniformity of energy distribution. Finally, all time-domain and frequency-domain feature parameters extracted from all sub-frequency band signals are combined. Specifically, these calculated values ​​can be concatenated in a predetermined order to form a single, high-dimensional numerical sequence, i.e., a multidimensional feature vector. For example, the root mean square value, kurtosis coefficient, impulse factor, main frequency band energy proportion, and frequency band energy entropy of all sub-frequency bands can be arranged sequentially to form a comprehensive feature vector. This multidimensional feature vector can then be used as vehicle feature vector information for subsequent crack condition diagnosis.

[0099] The above technical solution effectively overcomes the problem of inaccurate feature extraction caused by the presence of a large amount of noise and redundant information in the original vibration response signal during direct analysis. Through three-level wavelet packet decomposition, the vibration response signal is effectively denoised and decomposed into multiple sub-band signals with specific frequency ranges, highlighting the subtle features related to brake pad spring cracks. Furthermore, by comprehensively extracting the time-domain and frequency-domain feature parameters of each sub-band signal, key information such as signal energy, impact characteristics, waveform morphology, and frequency distribution can be captured from different dimensions, forming a more comprehensive and discriminative feature set. Finally, combining these multi-dimensional feature parameters into vehicle feature vector information significantly improves the representativeness and robustness of the feature vector, providing a more reliable and accurate basis for subsequent crack state judgment, thereby improving the diagnostic accuracy of brake pad spring crack states.

[0100] In some embodiments, the step of sending a synchronous trigger signal to an accelerometer simultaneously with the issuance of the standardized braking command, and the accelerometer acquiring the target vibration response signal generated by the brake pad spring under the standardized excitation at a preset high sampling rate, includes:

[0101] Repeat the standardized stimulus a preset number of times;

[0102] Each time the standardized excitation is applied, the vibration response signal generated by the brake pad spring under the standardized excitation is collected synchronously to obtain multiple vibration response signals;

[0103] Multiple vibration response signals are time-aligned and superimposed to suppress noise, thus obtaining the target vibration response signal.

[0104] This involves synchronizing or aligning multiple vibration response signals collected at different time points using a specific algorithm or technique. This time alignment can be achieved by employing cross-correlation algorithms, peak detection, or methods based on specific event markers. This ensures the phase consistency of the effective signals during subsequent superposition processing, thereby preventing signal cancellation.

[0105] The technique of arithmetically superimposing or averaging multiple time-aligned vibration response signals to enhance the effective components of the signal and reduce random noise achieves superposition noise suppression processing. Specifically, it can be implemented by methods such as coherent superposition or integrated averaging, which is used to improve the signal-to-noise ratio of the signal, thereby obtaining a clearer and more reliable target vibration response signal.

[0106] This application obtains multiple vibration response signals by repeatedly applying a standardized excitation a preset number of times. Compared to a single excitation, multiple excitations accumulate more effective signal energy, improving the signal-to-noise ratio. During each application of the standardized excitation, the vibration response signal generated by the brake spring under the standardized excitation is simultaneously acquired, resulting in multiple vibration response signals. Simultaneous acquisition ensures that each acquired signal precisely corresponds to the standardized excitation, providing a foundation for subsequent time alignment and superposition processing. The multiple vibration response signals are time-aligned and superimposed to suppress noise, yielding the target vibration response signal. Time alignment ensures the correspondence of each vibration response signal on the time axis, enabling superposition processing to effectively enhance common components in the signal while suppressing random noise. Superposition noise suppression utilizes the coherence of the signal, meaning that the effective signal has similar waveforms and phases in multiple measurements, while noise is random. Through superposition, the amplitude of the effective signal is enhanced, while the amplitude of the noise is reduced due to random cancellation, thereby improving the signal-to-noise ratio of the target vibration response signal.

[0107] The higher-quality target vibration response signal obtained through the above processing provides a cleaner data source for subsequent advanced feature extraction. When the target vibration response signal with a higher signal-to-noise ratio is subjected to three-level wavelet packet decomposition and the time-domain and frequency-domain feature parameters of each sub-band signal are extracted, these feature parameters will more accurately reflect the true vibration characteristics of the brake pad spring and reduce the interference of noise on feature calculation. Furthermore, these more accurate time-domain and frequency-domain feature parameters are combined into a multi-dimensional feature vector, enabling the vehicle feature vector information to more accurately characterize the crack state of the brake pad spring.

[0108] Finally, when the multidimensional feature vector similarity comparison algorithm is used to compare the vehicle feature vector information with the feature set in the preset standard sample library, the accuracy and reliability of the comparison results will be significantly improved. This will enable a more accurate determination of the true crack state of the brake pad spring, effectively avoid misjudgment or omission due to signal quality issues, and improve the robustness and diagnostic accuracy of the entire crack data analysis method.

[0109] For example, the preset number of times the standardized excitation is applied can be set to 10 to 20 times to control the detection time while ensuring signal enhancement. During each application of the standardized excitation, the vibration response signal generated by the brake pad spring under the standardized excitation is synchronously acquired. This can be achieved by the braking system control unit sending a synchronization pulse signal to the accelerometer simultaneously with the braking command. This pulse signal serves as the starting trigger point for accelerometer data acquisition, ensuring that each acquired vibration response signal is precisely aligned with the start time of the excitation. When aligning multiple vibration response signals in time, a method based on the cross-correlation function can be used. Specifically, the first acquired vibration response signal is selected as the reference signal, and then the cross-correlation function between each of the remaining vibration response signals and the reference signal is calculated. The peak position of the cross-correlation function is found, which represents the time delay. Based on this delay, the signal is shifted and compensated, thereby achieving time alignment of all signals. When performing superposition noise suppression processing, a coherent averaging method can be used. The time-aligned vibration response signals are arithmetically averaged point-by-point along the time axis. That is, for each time point, the amplitudes of all signals at that time point are summed and then divided by the number of signals. In this way, random noise components cancel each other out due to their randomness in multiple superpositions, while the effective vibration response signal generated by the normalized excitation is cumulatively enhanced due to its coherence, thus obtaining a target vibration response signal with a significantly improved signal-to-noise ratio.

[0110] Through the above technical solution, the problems that the vibration response signal generated by a single excitation is weak and vulnerable to noise interference, resulting in low signal quality, and further affecting the accuracy of subsequent feature extraction and crack state judgment are solved. By repeatedly applying standardized excitation and synchronously collecting multiple vibration response signals, and then performing time alignment and superposition to suppress noise, the effective signal components can be effectively enhanced, and the noise interference can be significantly reduced, so as to obtain the target vibration response signal with a higher signal-to-noise ratio. This enables more accurate and reliable vehicle feature vector information to be obtained when performing advanced feature extraction on the vibration response signal subsequently. Furthermore, through the multi-dimensional feature vector similarity comparison algorithm, the true crack state of the brake leaf spring can be determined more precisely, improving the accuracy and reliability of crack diagnosis.

[0111] In some embodiments, the feature sets in the preset standard sample library include a standard healthy brake leaf spring feature set and a standard cracked brake leaf spring feature set; the steps of comparing the vehicle feature vector information with the feature sets in the preset standard sample library by using the multi-dimensional feature vector similarity comparison algorithm to obtain a comparison result, and determining the true crack state of the brake leaf spring according to the comparison result specifically include:

[0112] Calculate the average distance DH between the vehicle feature vector information and all vectors in the standard healthy brake leaf spring feature set, and the average distance DC between the vehicle feature vector information and all vectors in the standard cracked brake leaf spring feature set;

[0113] If DC < DH and DC is less than a preset threshold, it is determined that the brake leaf spring is in a cracked state;

[0114] If DH < DC and DH is less than a preset threshold, it is determined that the brake leaf spring is in a healthy state.

[0115] Among them, the standard healthy brake leaf spring feature set refers to the set of feature vectors representing the healthy state constructed by collecting vibration response signals and performing advanced feature extraction on a large number of known healthy brake leaf springs. It can be constructed by means of clustering analysis, principal component analysis or expert experience annotation, etc., providing a clear healthy reference benchmark for subsequent crack state determination.

[0116] The standard cracked brake leaf spring feature set refers to the set of feature vectors representing the cracked state constructed by collecting vibration response signals and performing advanced feature extraction on a large number of known brake leaf springs with different degrees of cracks. It can be constructed by means of machine learning classification, anomaly detection or manual annotation, etc., and is used to provide a clear crack reference benchmark for subsequent crack state determination.

[0117] DH and DC can be measured by distance calculation methods such as Euclidean distance, Manhattan distance or cosine similarity.

[0118] The preset threshold is a critical value used to compare the average distance when judging the condition of the brake pad spring. It can be determined based on historical data analysis, expert experience, or optimization through machine learning model training. It can provide an adjustable judgment standard to balance the false alarm rate and the false alarm rate.

[0119] This application provides a clear classification basis for subsequent judgment by subdividing the preset standard sample library into a standard healthy brake pad spring feature set and a standard cracked brake pad spring feature set. After receiving the vehicle feature vector information obtained through advanced feature extraction, the system calculates the average distance DH between the vehicle feature vector information and all vectors in the standard healthy brake pad spring feature set, and the average distance DC between the vehicle feature vector information and all vectors in the standard cracked brake pad spring feature set. This method of calculating the average distance can effectively reduce the impact of a single abnormal sample or local noise on the overall comparison result, thereby enhancing the robustness of the comparison. Subsequently, the system performs a dual judgment based on the relative magnitude of DH and DC, combined with a preset threshold. Specifically, when the average distance DC of the cracked state is less than the average distance DH of the healthy state, and DC is also less than the preset threshold, the system determines that the brake pad spring is in a cracked state. This indicates that the vehicle feature vector information is closer to the cracked sample set, and the degree of closeness meets the preset crack standard. Conversely, when the average distance DH of the healthy state is less than the average distance DC of the cracked state, and DH is also less than a preset threshold, the system determines the brake pad spring to be in a healthy state. This indicates that the vehicle feature vector information is closer to the healthy sample set, and the degree of closeness meets the preset health standard. This judgment mechanism, which combines relative distance and absolute threshold, can effectively avoid misjudgments that may be caused by relying solely on a single similarity index, significantly improving the accuracy and reliability of brake pad spring crack state determination. Furthermore, this technical solution, combined with previous methods for obtaining vehicle feature vector information, can leverage even stronger advantages. Previous methods, by receiving contradictory multi-source data, issuing standardized braking commands, collecting vibration responses at high sampling rates, and performing advanced feature extraction, ensured that the vehicle feature vector information input to this judgment step was high-quality data that had undergone standardized excitation, denoising processing, and possessed rich discriminative capabilities. It is precisely based on this high-quality, high-reliability vehicle feature vector information that the judgment logic based on average distance and dual thresholds adopted in this technical solution can fully exert its effectiveness and accurately capture subtle state changes of the brake pad spring. This combination enables the system to not only overcome the limitations of traditional similarity comparison, but also effectively cope with noise and data deviations in practical applications. Thus, in the case of conflicting information from multiple sources, it provides a more accurate and reliable judgment of the true crack state of the brake pad spring, avoiding the continued operation of vehicles with safety hazards caused by misjudgment.

[0120] For example, this application is implemented as follows: First, a pre-defined standard sample library can be stored in a distributed database. The standard healthy brake pad spring feature set can be a collection of thousands of vibration feature vectors of healthy brake pad springs, obtained through standardized testing and feature extraction on brake pad springs of a large number of normally operating vehicles. The standard cracked brake pad spring feature set can be a collection of thousands of vibration feature vectors of brake pad springs known to have different types and degrees of cracks, obtained through testing and feature extraction on faulty vehicles or laboratory-simulated crack samples. When the system obtains the vehicle feature vector information of the vehicle to be analyzed, for example, if the vector is a 10-dimensional feature vector, the system will call a distance calculation module. This module can use Euclidean distance as the distance metric, calculate the Euclidean distance between the vehicle feature vector information and each vector in the standard healthy brake pad spring feature set, and then calculate the average of these distances to obtain the average distance DH. Simultaneously, this module will also calculate the Euclidean distance between the vehicle feature vector information and each vector in the standard cracked brake pad spring feature set, and calculate the average of these distances to obtain the average distance DC. For example, a preset threshold can be set to 0.5, obtained through cross-validation and optimization of a large amount of historical data. Subsequently, a decision logic unit receives DH, DC, and the preset threshold. If the calculated DC is 0.3, DH is 0.8, and the preset threshold is 0.5, since DC (0.3) is less than DH (0.8) and DC (0.3) is less than the preset threshold (0.5), the decision logic unit outputs that the brake spring is in a cracked state. Conversely, if DH is 0.2, DC is 0.7, and the preset threshold is 0.5, since DH (0.2) is less than DC (0.7) and DH (0.2) is less than the preset threshold (0.5), the decision logic unit outputs that the brake spring is in a healthy state. This specific implementation ensures the accuracy and reliability of the brake spring state determination.

[0121] Through the above technical solution, this application effectively solves the problems of insufficient accuracy and susceptibility to noise in the determination of brake pad spring crack status using traditional similarity comparison methods. By subdividing the standard sample library into two feature sets—healthy and cracked—and calculating the average distance between the feature vector information of the vehicle under test and these two feature sets, the overall attribution tendency of the sample under test can be more comprehensively measured. Combining the comparison of the relative magnitude of the average distance and the judgment of the absolute threshold forms a dual verification mechanism, which significantly enhances the accuracy and reliability of the judgment. This method can effectively distinguish between the healthy state and the cracked state of the brake pad spring, reduce the misjudgment rate caused by noise or data deviation, thereby ensuring the accurate identification of the true crack status of the brake pad spring and preventing vehicles with safety hazards from continuing to operate.

[0122] In some embodiments, after the step of receiving maintenance reports from third-party repair shops, vehicle vibration sensor data, and user feedback data:

[0123] The status of the maintenance report, the vehicle vibration sensor data, and the user feedback data is extracted to obtain the status indications corresponding to each data source.

[0124] Based on the type and historical performance of each data source, a reliability weight is assigned to the status indication;

[0125] Based on the status indication and the reliability weight, calculate the tendency score of each data source for the brake spring status;

[0126] By comparing the propensity scores, and identifying any discrepancies between the maintenance report, the vehicle vibration sensor data, and the user feedback data when there are preset differences between the propensity scores, inconsistencies are found.

[0127] This involves transforming data of different formats and types (such as structured maintenance reports, time-series vehicle vibration sensor data, and unstructured user feedback data) into unified and standardized state descriptions or numerical representations to achieve state extraction. This can be achieved using technologies such as natural language processing, signal processing, or data parsing, which is used to unify heterogeneous data and lay the foundation for subsequent quantitative comparisons.

[0128] Among them, the status indication refers to the standardized description or quantitative value of the current status of the brake pad spring by each data source after status extraction. It can be represented as a discrete category (such as "healthy", "abnormal", "to be observed") or a continuous value, which is used to provide a unified and comparable benchmark for brake pad spring status evaluation.

[0129] Among them, reliability weight refers to the trust level or influence coefficient assigned to the status indication based on the inherent characteristics, historical accuracy or data quality assessment results of each data source. It can be determined by means of expert experience assignment, historical data statistical analysis or dynamic adjustment by machine learning model, etc., to quantify the confidence level of different data sources, so as to give more reliable data greater influence in the judgment of contradictions.

[0130] Among them, the propensity score refers to the quantitative value of the brake spring status judgment of each data source after comprehensively considering the status indication and its reliability weight. It can be calculated by weighted summation, weighted average or other aggregation algorithms, and is used to integrate the judgments and credibility of each data source into a single, directly comparable value.

[0131] Among them, the preset difference refers to the threshold or set of rules used to judge whether there is a contradiction between the tendency scores. It can be determined according to the actual application scenario, experience value or through training with historical data. It is used to provide a judgment standard to avoid misjudging contradictions due to small fluctuations, while ensuring sensitivity to real contradictions.

[0132] This application effectively solves the challenge of identifying contradictions between heterogeneous data through a series of refined processing steps. First, it extracts the status of maintenance reports, vehicle vibration sensor data, and user feedback data, transforming the originally inconsistent and semantically ambiguous data into standardized status indicators. For example, textual descriptions in maintenance reports, waveform characteristics of sensor data, and natural language in user feedback are all mapped to unified status descriptions such as "healthy," "abnormal," or "performance degradation." This transformation forms the basis for subsequent quantitative analysis, ensuring the comparability of different data sources. Based on this, reliability weights are assigned to these status indicators according to the type and historical performance of each data source. Considering the inherent differences in accuracy and historical performance among different data sources—for example, rigorously calibrated sensor data is generally more objective than subjective user feedback, and reports from large professional repair shops may be more reliable than those from smaller repair shops—by assigning weights, the system can intelligently assess the confidence level of each data source, ensuring that more reliable information has a greater influence in subsequent judgments, thereby reducing the risk of misleading information from low-quality data. Subsequently, based on the obtained status indicators and reliability weights, the system calculates the bias score for each data source regarding the brake spring status. This step quantifies the status indication and performs a weighted calculation based on its reliability weights to obtain a numerical value that intuitively reflects the "tendency" of each data source in judging the health status of the brake spring. For example, a high-reliability "healthy" indication will generate a high positive score, while a high-reliability "abnormal" indication will generate a low negative score. This quantification process allows the judgments from different data sources to be compared on a unified numerical dimension. Ultimately, by comparing these tendency scores, the system can identify contradictions between maintenance reports, vehicle vibration sensor data, and user feedback data when preset differences exist. This comparison mechanism can capture potential inconsistencies between data sources; for example, when maintenance reports tend to be "healthy" while sensor data or user feedback strongly tends to be "abnormal," the system can accurately determine the existence of a contradiction. By setting preset differences, the sensitivity of contradiction identification can be flexibly adjusted to avoid frequent triggering due to minor fluctuations, while ensuring timely response to real safety hazards. Through the above-mentioned refined contradiction identification mechanism, this application can accurately judge the inconsistencies between multi-source heterogeneous data, thereby providing accurate triggering conditions for subsequent standardized braking command issuance. This means that the system will only initiate the time-consuming and resource-intensive standardized excitation and data acquisition process when there are genuine data discrepancies and further verification of the true state of the brake pad spring is required. This precise triggering mechanism avoids unnecessary testing, improves diagnostic efficiency, and ensures that the judgment of the true crack state of the brake pad spring is based on more reliable and comprehensive information, significantly improving the accuracy and robustness of the entire brake pad spring crack data analysis method.

[0133] In some embodiments, the step of assigning reliability weights to the status indication based on the type and historical performance of each data source includes:

[0134] Based on the type and historical performance of each data source, an initial reliability weight is assigned to the status indication;

[0135] Continuously monitor the consistency between the status indications of each data source and the actual crack state of the brake pad spring;

[0136] Based on the consistency, the initial reliability weight is adjusted to obtain the adjusted reliability weight, and the adjusted reliability weight is used as the reliability weight of the state indication.

[0137] The initial reliability weight refers to the preliminary assessment of the reliability of data sources based on their type and historical performance. It can be set by expert experience, statistical analysis based on historical datasets, or a pre-set rule base, and is used to provide a benchmark for subsequent dynamic adjustments.

[0138] The system compares the consistency between the status indications given by each data source and the actual crack state of the brake pad spring in real time or periodically, and continuously monitors the consistency between the status indications of each data source and the actual crack state of the brake pad spring. This can be achieved by comparing the data source indications with manual inspection results, results from higher-level diagnostic systems, or subsequent actual failure occurrences. This is used to obtain the performance data of the data source in actual operation and to provide a basis for weight adjustment.

[0139] Based on the monitored consistency data, the preset initial reliability weights are dynamically adjusted. This can be achieved using machine learning algorithms, adaptive filtering algorithms, or rule-based iterative update mechanisms. The aim is to enable the weights to adapt to the real-time performance of the data source and environmental changes, thereby improving the accuracy of the assessment.

[0140] This application addresses the problem of inaccurate assessments that may result from relying solely on initial weights by introducing a mechanism for dynamically adjusting reliability weights. Specifically, the system first assigns an initial reliability weight to the status indications of each data source based on its type and historical performance, such as the professionalism of maintenance reports, the objectivity of vehicle vibration sensor data, and the subjectivity of user feedback data. This initial weight reflects the judgment of the prior credibility of the data source. However, considering that the actual performance of the data source may change or deviate over time, this technical solution further continuously monitors the consistency between the status indications of each data source and the actual crack state of the brake pad spring. This means that the system will continuously compare the status reported by the data source with the actual crack state obtained through other more reliable means (such as actual inspection or subsequent fault confirmation). When it is found that the indication of a data source is inconsistent with the actual state for a long period of time, the system will adjust the initial reliability weight of that data source based on this inconsistency. For example, if a maintenance report from a repair shop repeatedly does not match the actual crack state, its weight will be reduced; conversely, if the indication of a data source consistently matches the actual state highly, its weight will be increased. Through continuous monitoring and dynamic adjustment, the system obtains adjusted reliability weights, which more accurately reflect the current credibility of the data sources. These adjusted reliability weights are then used to calculate the bias scores of each data source regarding the brake pad spring condition. This allows for more accurate judgments based on weights when identifying contradictions between maintenance reports, vehicle vibration sensor data, and user feedback data, significantly improving the accuracy of contradiction identification and the reliability of determining the true crack condition of the brake pad spring. It is precisely because of this dynamic learning and adaptability that this technical solution effectively overcomes the limitations of single or fixed weight evaluation, ensuring that the system can still make accurate judgments even when there are contradictions in multi-source heterogeneous information.

[0141] For example, this application is implemented as follows: The system first assigns an initial reliability weight to the status indication based on the type and historical performance of each data source. For instance, maintenance reports from authoritative third-party repair shops can be assigned a higher initial weight, such as 0.8; vehicle vibration sensor data, due to its objectivity, can be assigned a medium weight, such as 0.6; and user feedback data, due to its subjectivity, can be assigned a lower initial weight, such as 0.4. Subsequently, the system continuously monitors the consistency between the status indications of each data source and the actual crack state of the brake pad spring. This can be achieved by establishing a feedback loop. For example, after the vehicle undergoes manual inspection or more precise diagnostic equipment confirms the actual crack state, the system compares the actual state with the status indications previously given by each data source. If the indication from a certain data source does not match the actual state, for example, the maintenance report shows "no cracks" but the actual inspection finds "cracks," the system records this inconsistency. Based on this consistency or inconsistency, the system dynamically adjusts the initial reliability weight. For example, an iterative update algorithm can be used. If the data source's indication matches the true state, its weight will increase slightly; if the data source's indication does not match the true state, its weight will decrease accordingly. After multiple iterations and learning, the system will obtain adjusted reliability weights, which can more accurately reflect the actual credibility of the current data source. Finally, the system uses these adjusted reliability weights as the reliability weights for the state indication, which are used for subsequent propensity score calculation and contradiction identification, thereby improving the accuracy of brake pad spring crack state determination.

[0142] In some embodiments, the step of extracting the status of the maintenance report, the vehicle vibration sensor data, and the user feedback data to obtain the status indication corresponding to each data source includes:

[0143] The maintenance report is subjected to structured information parsing and keyword extraction, and the parsing results are mapped to the status indication of the brake pad spring;

[0144] The time-domain or frequency-domain features of the vehicle vibration sensor data are extracted, and the extracted features are compared with a preset feature pattern to obtain the state indication of the brake pad spring.

[0145] In addition, natural language processing is performed on the user feedback data to identify keywords or phrases related to the state of the brake pad spring, and the identification results are mapped to the state indication of the brake pad spring.

[0146] This involves identifying, extracting, and classifying data with predefined formats or fields to achieve structured information parsing. This can be achieved using regular expression matching, template parsing, or entity recognition models based on machine learning.

[0147] Keyword extraction is the process of identifying words or phrases that represent the core content of unstructured text. It can be achieved using word frequency statistics (such as TF-IDF), graph models (such as TextRank), or deep learning-based sequence labeling models.

[0148] Extracting parameters that characterize the properties of a signal in the time or frequency domain from the original signal, and extracting time or frequency domain features, can be achieved through statistical analysis (such as mean, variance, peak value), transform analysis (such as Fourier transform, wavelet transform), or feature learning methods based on deep learning.

[0149] Preset feature patterns refer to a set of typical features that are predefined based on historical data or expert experience before data analysis, and are used for comparison and classification. These features can take the form of cluster centers, classifier model parameters, or expert rule bases.

[0150] This application employs customized data extraction methods tailored to the characteristics of different data sources, thereby accurately and comprehensively extracting effective information reflecting the true state of brake pad springs from heterogeneous data sources. Specifically, since maintenance reports typically contain structured inspection data and unstructured descriptions from maintenance personnel, this application performs structured information parsing on the maintenance reports to accurately extract quantitative indicators such as wear level and replacement date. Simultaneously, through keyword extraction, it can identify key descriptions such as "cracks" and "abnormal wear" from free text. The combination of these two methods allows for comprehensive and detailed capture of information in the maintenance reports, uniformly mapped to brake pad spring status indicators. Given that vehicle vibration sensor data can directly reflect changes in the physical characteristics of brake pad springs, this application extracts time-domain or frequency-domain features to transform the original vibration signal into a quantifiable feature vector, such as energy distribution and frequency components. Subsequently, these extracted features are compared with preset feature patterns of healthy or cracked brake pad springs to determine whether the current vibration state of the brake pad spring is abnormal and obtain corresponding status indicators. This feature comparison-based method can effectively identify subtle vibration pattern changes in brake pad springs caused by cracks. Furthermore, considering that user feedback data, as an unstructured subjective description, can provide intuitive feelings and abnormal signs that are difficult for sensors and maintenance reports to capture, this application employs natural language processing technology to identify keywords or phrases related to the brake spring condition, such as "soft brakes" and "abnormal noise," from user feedback text and map them as condition indicators. This processing method overcomes the challenge of unstructured user feedback data, transforming it into analyzable state information. Because targeted and refined state extraction methods are used for maintenance reports, vehicle vibration sensor data, and user feedback data, more accurate, comprehensive, and reliable brake spring condition indicators can be obtained from these heterogeneous and potentially contradictory data sources. These refined state indicators serve as the basis for subsequent identification of data contradictions, providing high-quality input for determining whether there are pre-set discrepancies between maintenance reports, vehicle vibration sensor data, and user feedback data. Based on this, by combining the reliability weight allocation for each data source type and historical performance, and adjusting the reliability weight by continuously monitoring the consistency between the condition indicators and the actual crack condition, contradictions between data can be identified more accurately, providing solid data support for judging the actual crack condition of the brake spring.

[0151] In some embodiments, the step of calculating the tendency score of each data source for the brake spring state includes:

[0152] The status indicators of each data source are converted into numerical values, and the numerical values ​​are weighted and summed or weighted averaged according to the reliability weights to obtain the propensity score.

[0153] This involves quantifying the state information expressed by different data sources into a numerical form that can be used for mathematical operations. This can be achieved using preset mapping rules, encoding tables, or data processing algorithms, providing a unified data foundation for subsequent comprehensive calculations.

[0154] The converted values ​​are multiplied by the corresponding reliability weights and then summed, or the weighted values ​​are averaged to achieve weighted summation or weighted average. This can be achieved using linear weighting, exponential weighting, or other non-linear weighting algorithms, so that more reliable data sources contribute more to the final result, thereby reducing the negative impact of unreliable data sources.

[0155] This application eliminates the differences between different data types by uniformly quantifying the status indications from heterogeneous data sources such as maintenance reports, vehicle vibration sensors, and user feedback, enabling quantitative comparison and calculation of this information. Based on this, these values ​​are weighted and summed or averaged according to predetermined or dynamically adjusted reliability weights. This ensures that data sources with more reliable historical performance and higher data quality have a greater weight in the final propensity score calculation, while the influence of less reliable data sources is effectively weakened. It is precisely this quantification and weighting process that allows the propensity scores of each data source for brake spring status to more accurately reflect their true state and effectively reduces the interference of a single unreliable data source on the overall judgment. The propensity scores obtained in this way can then be used for comparison. When there are preset differences between these propensity scores, the system can more accurately identify contradictions between maintenance reports, vehicle vibration sensor data, and user feedback data, thereby triggering further detection of the brake spring status. This processing mechanism ensures that even when faced with conflicting and inconsistent data, the system can make judgments based on more comprehensive and weighted information, significantly improving the accuracy of contradiction identification and the reliability of subsequent crack condition assessment.

[0156] On the other hand, such as Figure 2 As shown, an exemplary brake pad spring crack data analysis system is illustrated. This application further proposes a brake pad spring crack data analysis system 100, which includes:

[0157] The data receiving module 10 is used to receive maintenance reports from third-party repair stations, vehicle vibration sensor data, and user feedback data.

[0158] The excitation control module 20 is used to issue a standardized braking command to the braking system of the target vehicle when a contradiction is detected between the maintenance report, the vehicle vibration sensor data, and the user feedback data, and the vehicle engine is off. This command drives the brake caliper to perform a clamping action, causing the brake pad spring to receive standardized excitation and generate a vibration response. The brake caliper clamps the brake pad assembly, which includes a brake pad spring and brake friction pads.

[0159] The vibration response acquisition module 30 is used to send a synchronous trigger signal to the accelerometer at the same time as the standardized braking command is issued. The accelerometer acquires the target vibration response signal generated by the brake pad spring under the standardized excitation at a preset high sampling rate.

[0160] Feature extraction module 40: Performs advanced feature extraction on the target vibration response signal to obtain vehicle feature vector information;

[0161] The crack diagnosis module 50 uses a multi-dimensional feature vector similarity comparison algorithm to compare the vehicle feature vector information with the feature sets in the preset crack sample library and health sample library. When the similarity with the crack sample library exceeds a first threshold, a crack diagnosis conclusion is output; when the similarity with the health sample library exceeds a second threshold, a health diagnosis conclusion is output.

[0162] Through the above technical solution, this application provides a brake pad spring crack data analysis system, solving the problem in the prior art where inconsistencies exist between maintenance reports, vehicle vibration sensor data, and user feedback data, making it difficult to determine the condition of the brake pad spring. This system integrates multi-source heterogeneous data through a data receiving module. When a data inconsistency is detected, the excitation control module actively issues a standardized braking command to the braking system, driving the brake caliper to perform a clamping action, causing the brake pad spring to receive standardized excitation and generate a vibration response. The vibration response acquisition module synchronously acquires the target vibration response signal at a high sampling rate, ensuring data reliability. The feature extraction module performs advanced feature extraction on the signal to obtain vehicle feature vector information, reducing data dimensionality while retaining core diagnostic information. The crack diagnosis module uses a multi-dimensional feature vector similarity comparison algorithm to compare the vehicle feature vector information with a preset crack sample library and a healthy sample library, thereby outputting a crack or health diagnosis conclusion for the brake pad spring. The collaborative work of these modules enables the system to obtain diagnostic evidence through proactive and standardized stimulus and response collection when faced with complex and contradictory raw data. This improves the accuracy of brake pad spring crack diagnosis and avoids safety hazards caused by misjudgment.

[0163] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for analyzing brake pad spring crack data, characterized in that, include: Receive maintenance reports from third-party repair shops, vehicle vibration sensor data, and user feedback data; When a discrepancy is detected between the maintenance report, the vehicle vibration sensor data, and the user feedback data, and the vehicle engine is off, a standardized braking command is issued to the target vehicle's braking system to drive the brake caliper to perform a clamping action, causing the brake pad spring to receive standardized excitation and generate a vibration response; wherein, the brake caliper clamps the brake pad assembly, and the brake pad assembly includes the brake pad spring and the brake friction pad; Simultaneously with the issuance of the standardized braking command, a synchronous trigger signal is sent to the accelerometer, which acquires the target vibration response signal generated by the brake pad spring under the standardized excitation at a preset high sampling rate; Advanced feature extraction is performed on the target vibration response signal to obtain vehicle feature vector information; A multidimensional feature vector similarity comparison algorithm is used to compare the vehicle feature vector information with the feature set in a preset standard sample library to obtain the comparison result, and the actual crack state of the brake pad spring is determined based on the comparison result. Following the step of receiving maintenance reports from third-party repair shops, vehicle vibration sensor data, and user feedback data, the following is included: The status of the maintenance report, the vehicle vibration sensor data, and the user feedback data is extracted to obtain the status indications corresponding to each data source. Based on the type and historical performance of each data source, a reliability weight is assigned to the status indication; Based on the status indication and the reliability weight, calculate the tendency score of each data source for the brake spring status; By comparing the propensity scores, and identifying any discrepancies between the maintenance report, the vehicle vibration sensor data, and the user feedback data when there are preset differences between the propensity scores, inconsistencies are found.

2. The brake pad spring crack data analysis method according to claim 1, characterized in that, Before the step of issuing a standardized braking command to the braking system of the target vehicle to drive the brake calipers to perform a clamping action, the following steps are also included: A standardized physical surface pretreatment command is issued to the braking system, driving the brake caliper to perform multiple clamping and releasing operations on the brake pad spring according to the set pressure and duration; wherein, the clamping and releasing operations cause the brake friction pads in the brake pad assembly to rub against the brake disc, so as to remove the unstable oxide film on the friction surface.

3. The method for analyzing brake pad spring crack data according to claim 1, characterized in that, The step of performing high-level feature extraction on the target vibration response signal to obtain vehicle feature vector information includes: The target vibration response signal is subjected to three-level wavelet packet decomposition to obtain sub-frequency band signals; Extract the time-domain and frequency-domain feature parameters of each sub-band signal, wherein: the time-domain feature parameters include: root mean square value, kurtosis coefficient, and impulse factor; the frequency-domain feature parameters include: main band energy proportion and band energy entropy; The time-domain feature parameters and frequency-domain feature parameters are combined into a multi-dimensional feature vector, which serves as the vehicle feature vector information.

4. The method for analyzing brake pad spring crack data according to claim 1, characterized in that, While the standardized braking instruction is issued, a synchronous trigger signal is sent to the accelerometer. The steps for the accelerometer to collect the target vibration response signal generated by the brake leaf spring under the standardized excitation at a preset high sampling rate include: Repeatedly applying the standardized excitation a preset number of times; During each application of the standardized excitation, synchronously collecting the vibration response signal generated by the brake leaf spring under the standardized excitation to obtain multiple vibration response signals; Performing time alignment on the multiple vibration response signals and performing superposition to suppress noise to obtain the target vibration response signal.

5. The method for analyzing brake pad spring crack data according to claim 1, characterized in that, The feature sets in the preset standard sample library include the standard healthy brake leaf spring feature set and the standard cracked brake leaf spring feature set; The steps for comparing the vehicle feature vector information with the feature sets in the preset standard sample library using the multi-dimensional feature vector similarity comparison algorithm to obtain the comparison result and determining the true crack state of the brake leaf spring according to the comparison result specifically include: Calculating the average distance DH between the vehicle feature vector information and all vectors in the standard healthy brake leaf spring feature set, and the average distance DC between the vehicle feature vector information and all vectors in the standard cracked brake leaf spring feature set; If DC < DH and DC is less than the preset threshold, it is determined that the brake leaf spring is in a cracked state; If DH < DC and DH is less than the preset threshold, it is determined that the brake leaf spring is in a healthy state.

6. The method for analyzing brake pad spring crack data according to claim 1, characterized in that, The steps for assigning reliability weights to the status indication according to the types and historical performances of the respective data sources include: Assigning initial reliability weights to the status indication according to the types and historical performances of the respective data sources; Continuously monitoring the consistency between the status indication of each data source and the true crack state of the brake leaf spring; Adjusting the initial reliability weight according to the consistency to obtain the adjusted reliability weight, and using the adjusted reliability weight as the reliability weight of the status indication.

7. The method for analyzing brake pad spring crack data according to claim 1, characterized in that, The steps for extracting the status from the maintenance report, the in-vehicle vibration sensor data, and the user feedback data to obtain the status indication corresponding to each data source include: Performing structured information parsing and keyword extraction on the maintenance report and mapping the parsing result to the status indication of the brake leaf spring; Performing time-domain or frequency-domain feature extraction on the in-vehicle vibration sensor data and comparing the extracted features with the preset feature patterns to obtain the status indication of the brake leaf spring; And performing natural language processing on the user feedback data to identify keywords or phrases related to the status of the brake leaf spring and mapping the identification result to the status indication of the brake leaf spring.

8. The method for analyzing brake pad spring crack data according to claim 1, characterized in that, The steps for calculating the tendency score of each data source for the status of the brake leaf spring include: Converting the status indication of each data source into a numerical value and performing weighted summation or weighted averaging on the numerical value according to the reliability weight to obtain the tendency score.

9. A brake pad spring crack data analysis system, characterized in that, The system includes: A data receiving module for receiving the maintenance report, the in-vehicle vibration sensor data, and the user feedback data from a third-party repair station; An excitation control module is used to issue a standardized braking command to the braking system of the target vehicle when a contradiction is detected between the maintenance report, the vehicle vibration sensor data, and the user feedback data, and the vehicle engine is off. This command drives the brake caliper to perform a clamping action, causing the brake pad spring to receive standardized excitation and generate a vibration response. The brake caliper clamps the brake pad assembly, which includes a brake pad spring and brake friction pads. The vibration response acquisition module is used to send a synchronous trigger signal to the accelerometer at the same time as the standardized braking command is issued. The accelerometer acquires the target vibration response signal generated by the brake pad spring under the standardized excitation at a preset high sampling rate. Feature extraction module: performs advanced feature extraction on the target vibration response signal to obtain vehicle feature vector information; The crack diagnosis module uses a multi-dimensional feature vector similarity comparison algorithm to compare the vehicle feature vector information with the feature sets in the preset crack sample library and health sample library. When the similarity with the crack sample library exceeds a first threshold, a crack diagnosis conclusion is output; when the similarity with the health sample library exceeds a second threshold, a health diagnosis conclusion is output. It is also used to extract the status of the maintenance report, the vehicle vibration sensor data and the user feedback data to obtain the status indication corresponding to each data source; Based on the type and historical performance of each data source, a reliability weight is assigned to the status indication; Based on the status indication and the reliability weight, calculate the tendency score of each data source for the brake spring status; By comparing the propensity scores, and identifying any discrepancies between the maintenance report, the vehicle vibration sensor data, and the user feedback data when there are preset differences between the propensity scores, inconsistencies are found.

Images