A LightGBM-based method for accurately locating automobile tire pressure faults
The LightGBM model and drools rule algorithm accurately locate the car tire pressure faults, which solves the problem of insufficient data in traditional diagnosis, improves detection efficiency and accuracy, and achieves real-time fault feedback.
Patent Information
- Application Number
- CN202210782304.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-07-05
- Publication Date
- 2025-08-26
- Estimated Expiration
- 2042-07-05
AI Technical Summary
In traditional automobile fault diagnosis, insufficient data and limited diagnostic information are found, difficult to reproduce and locate, resulting in inefficient maintenance and serious waste of resources.
The LightGBM model is used combined with the drools rule algorithm, and the vehicle tire pressure data is preprocessed and feature fusion is used to train accurate classification models, set threshold optimization models, and improve the accuracy and recall rate of fault detection.
It realizes the precise positioning of car tire pressure failures, improves detection efficiency and accuracy, reduces resource waste, and provides real-time feedback.
Smart Images
Figure CN115130517B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of computer automobile fault detection, and specifically to a method for accurately locating automobile tire pressure faults based on LightGBM. Background Art
[0002] With the increasing popularity of intelligent connected vehicles and the gradual advancement of autonomous driving-related functions, the application of AI algorithms in the automotive industry is becoming increasingly widespread. AI applications within vehicle manufacturers are primarily concentrated in two areas: first, autonomous driving, primarily encompassing environmental perception recognition algorithms, multi-channel perception fusion algorithms, and autonomous driving decision-making algorithms. Second, intelligent customer services, primarily encompassing speech recognition algorithms, NLP analysis algorithms, intelligent business recommendation algorithms, and business diagnosis algorithms.
[0003] To master core technologies and achieve competitive differentiation, automakers are gradually building their own AI capabilities. With the gradual expansion of AI teams and the increasing number of AI models, automakers are urgently seeking to build a unified big data model platform tailored to their needs. This platform can handle the unified access, preprocessing, labeling, training, and model hosting of machine learning data, improving modeling efficiency, unifying modeling methods, and reducing resource waste.
[0004] In traditional business, vehicle failures have always been an area that automakers devote significant effort to addressing during the R&D and after-sales stages. At the very least, these issues can lead to a decline in user experience and reputation, while at worst, they can threaten driving safety. Severe batch failures can even trigger large-scale vehicle recalls. Traditional vehicle repair services suffer from operational drawbacks such as insufficient diagnostic data, limited diagnostic information, and difficulty in 4S dealership repairs. Occasional failures are difficult to reproduce, resulting in lengthy troubleshooting and difficulty locating the problem. Consequently, vehicle repairs often suffer from low first-time resolution rates, repeated part replacements, and frequent, incorrect replacements. Therefore, the precise fault location algorithm developed and researched in this project and its application to the automotive industry is crucial. Summary of the Invention
[0005] The present invention aims to provide a method for accurately locating automobile tire pressure faults based on LightGBM to solve the problem of predicting and determining automobile tire faults.
[0006] In order to achieve the above objectives, the basic scheme of the present invention is as follows: a method for accurately locating automobile tire pressure failure based on LightGBM,
[0007] The steps include:
[0008] Step 1: Collect vehicle-side tire pressure data and tire pressure fault data respectively, preprocess the two sets of data, separate and fuse the preprocessed signal data to obtain tire pressure abnormality signal feature data; use the Drools rule algorithm to determine the tire pressure fault and obtain the fault label;
[0009] Step 2: Based on the LightGBM model, the tire pressure abnormality signal feature data and the corresponding fault labels are trained to obtain an accurate classification model for the fault abnormality type;
[0010] Step 3: Split the feature data into a training set and a test set, use the test set to test the trained accurate classification model, and set a preset threshold. If the accuracy of the accurate classification model test is greater than the preset threshold, the establishment of the accurate classification model is completed. If the accuracy of the trained accurate classification model test is less than or equal to the preset threshold, the drools rule algorithm is used to judge the training error data and increase the test accuracy to greater than the preset threshold.
[0011] Principle of the basic scheme: The device first modifies the LightGBM model based on the weak classification module decision tree algorithm, adjusts the parameters in the hierarchical structure, and improves it into a model suitable for accurate multi-classification of large amounts of data. It focuses on optimizing the problems of high density of automobile signal acquisition, large amount of data, and a large amount of redundancy in the collected data. The final model greatly improves the precision, recall and efficiency of automobile tire pressure fault detection, and provides conditions for accurate positioning and real-time detection of automobile tire pressure faults.
[0012] Advantages of the basic solution: Directly process and classify the data collected by the vehicle side, and make full use of the detailed information of the relevant signals collected by the vehicle side.
[0013] Furthermore, the specific method for extracting characteristic data of abnormal tire pressure signals is to collect signals of time, tire pressure value, ambient temperature, plateau coefficient and power status feedback of each tire on the vehicle side before and after the abnormality occurs to obtain raw data, clean the raw data, eliminate duplicate values and fill in missing values, and then convert the data type of the raw data to obtain a vehicle condition data table; use the drools rule algorithm to judge the tire pressure fault and obtain a fault label; merge the fault label with the vehicle condition data table, reduce the dimension through principal component analysis, and then remove signal features that have little impact on the occurrence of the fault to complete the feature combination.
[0014] Advantages of the basic solution: By leveraging the efficient parallel characteristics of the LightGBM model, the time cost of training the model is reduced, and real-time feedback on vehicle-side faults can be provided.
[0015] Furthermore, the specific method of the training is: merging the obtained vehicle condition data tables, compressing the data before and after one hour into one piece of data to provide data support for training; splitting the data into training sets and test sets with a ratio of 80% and 20%, extracting classification labels, and the classification labels are divided into air leakage, low tire pressure and normal. By training the abnormal tire pressure signal feature data and the corresponding classification labels, the accuracy and recall rate of fault prediction are improved.
[0016] Advantages of the basic solution: It provides scientific data support for the precise positioning of tire pressure failures. Through reasonable classification labels, the positioning results are fast and accurate, making it easier for drivers to carry out the next step of repairing the failure.
[0017] Additional aspects and advantages of the present invention will be set forth in part in the description which follows and, in part, will be obvious from the description which follows, or may be learned by practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS
[0018] Figure 1 It is a flow chart of a specific solution of the present invention.
[0019] Figure 2 It is the accuracy of the present invention in predicting air leakage failure.
[0020] Figure 3 It is the accuracy of the present invention in predicting tire pressure failure. DETAILED DESCRIPTION
[0021] The following describes embodiments of the present invention in detail. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended only to explain the present invention and are not to be construed as limiting the present invention.
[0022] In the description of the present invention, it should be understood that the terms "longitudinal", "transverse", "vertical", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating the orientation or position relationship, are based on the orientation or position relationship shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention.
[0023] In the description of the present invention, unless otherwise specified and limited, it should be noted that the terms "installed", "connected" and "connected" should be understood in a broad sense. For example, it can be a mechanical connection or an electrical connection, or it can be the internal communication between two components. It can be a direct connection or an indirect connection through an intermediate medium. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to the specific circumstances.
[0024] The following is further described in detail through specific implementation methods:
[0025] The embodiment is basically as shown in the attached Figure 1 , Attachment Figure 2 and attached Figure 3 As shown: A method for accurately locating automobile tire pressure failure based on LightGBM, including collecting vehicle-side tire pressure data and tire pressure failure data respectively, preprocessing the two sets of data to obtain preprocessed signal data, separating and feature-fusing the preprocessed signal data, and obtaining tire pressure abnormality signal feature data; using the drools rule algorithm to judge the vehicle tire pressure failure to obtain a fault label, and then based on the LightGBM model, training the tire pressure abnormality signal feature data and the corresponding fault label to obtain an accurate classification model of the fault abnormality type; splitting the feature data into a training set and a test set, using the test set to test the trained accurate classification model, and setting a preset threshold. If the accuracy of the accurate classification model test is greater than the preset threshold, the establishment of the accurate classification model is completed. If the accuracy of the trained accurate classification model test is less than or equal to the preset threshold, the drools rule algorithm is used to judge the training error data, and the test accuracy is increased to greater than the preset threshold.
[0026] The specific implementation process is as follows: The device first uses the LightGBM (Light Global Blockchain Model) decision tree algorithm as its foundation. The model is modified and the parameters within the hierarchical structure are adjusted to create a model suitable for accurate multi-classification of large amounts of data. This model specifically addresses the high density of vehicle signal acquisition, the large amount of data, and the significant redundancy in collected data. The resulting model significantly improves the precision, recall, and efficiency of tire pressure fault detection, facilitating the precise location and real-time detection of tire pressure faults.
[0027] As attached Figure 1As shown: The specific method for extracting characteristic data of abnormal tire pressure signals is to collect the time, tire pressure value, ambient temperature, plateau coefficient and power supply status feedback signals of each tire on the vehicle side before and after the abnormality occurs to obtain the original data, clean the original data, eliminate duplicate values and fill in missing values, and then convert the data type of the original data to obtain a vehicle condition data table; use the drools rule algorithm to judge the tire pressure fault and obtain a fault label; merge the fault label with the vehicle condition data table, reduce the dimension through principal component analysis, and then remove the signal features that have little impact on the occurrence of the fault to complete the feature combination.
[0028] The specific implementation process is as follows: Through the precise collection of various tire data, the data is cleaned, eliminated and filled. After the data type conversion, the vehicle condition data reflects the real-time fault status of the car tire, so that the final model can improve the precision, completeness and efficiency of automobile tire pressure fault detection, and provide accurate data for the precise positioning and real-time detection of automobile tire pressure faults.
[0029] As attached Figure 1 As shown: The specific training method is as follows: merging the obtained vehicle condition data tables, compressing the data before and after one hour into one data to provide data support for training; splitting the data into a training set and a test set with a ratio of 80% and 20%, extracting classification labels, and the classification labels are divided into air leakage, low tire pressure and normal. By training the abnormal tire pressure signal feature data and the corresponding classification labels, the accuracy and recall rate of fault prediction are improved.
[0030] The specific implementation process is as follows: Provide scientific data support for the precise positioning of automobile tire pressure failures, and through reasonable classification labels, make the positioning results fast and accurate, making it convenient for drivers to carry out the next step of repairing the failure.
[0031] As attached Figure 2 ,The accuracy of the leakage fault diagnosis result is 98.8%.,The y-axis of the confusion matrix is the true label, the horizontal axis is the predicted label.,Both use 0 to
[0032] As attached Figure 3 As shown in the figure, the accuracy of the tire pressure fault diagnosis result is 94.5%. The vertical axis of the confusion matrix is the true label, and the horizontal axis is the predicted label. 0 is used to represent no fault and 1 is used to represent fault.
[0033] The above is only an embodiment of the present invention, and common knowledge such as the specific structure and / or characteristics of the scheme is not described in detail here. It should be pointed out that for those skilled in the art, without departing from the structure of the present invention, several variations and improvements can be made, which should also be regarded as the scope of protection of the present invention, and these will not affect the effect of the implementation of the present invention and the practicality of the patent. The scope of protection required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to interpret the content of the claims.
Claims
1. A LightGBM-based method for accurately locating automobile tire pressure faults, characterized by: The steps include: Step 1: Collect vehicle-side tire pressure data and tire pressure fault data respectively, preprocess the two sets of data, separate and fuse the preprocessed signal data, and obtain tire pressure abnormality signal feature data; Use the drools rule algorithm to determine the tire pressure fault and obtain the fault label; Step 2: Based on the LightGBM model, the tire pressure abnormality signal feature data and the corresponding fault labels are trained to obtain an accurate classification model for the fault abnormality type; Step 3: Split the feature data into a training set and a test set, use the test set to test the trained accurate classification model, and set a preset threshold. If the accuracy of the accurate classification model test is greater than the preset threshold, the establishment of the accurate classification model is completed. If the accuracy of the trained accurate classification model test is less than or equal to the preset threshold, the drools rule algorithm is used to judge the training error data and increase the test accuracy to greater than the preset threshold. The specific method for extracting characteristic data of tire pressure abnormality signals is as follows: collecting signals of the time before and after the abnormality occurs on each tire on the vehicle side, tire pressure value, ambient temperature, plateau coefficient, and power supply status feedback to obtain raw data; cleaning the raw data, eliminating duplicate values, and filling in missing values; and then converting the data type of the raw data to obtain a vehicle condition data table; determining the tire pressure fault using the Drools rule algorithm to obtain a fault label; merging the fault label with the vehicle condition data table, reducing the dimension through principal component analysis, and then removing signal features that have little impact on the fault to complete feature combination; The specific training method is as follows: merging the obtained vehicle condition data tables, compressing the data before and after one hour into one data piece to provide data support for training; splitting the data into a training set and a test set with a ratio of 80% and 20%, extracting classification labels, and the classification labels are divided into air leakage, low tire pressure and normal. By training the abnormal tire pressure signal feature data and the corresponding classification labels, the accuracy and recall rate of fault prediction are improved.