Method and apparatus for predicting tire wear using machine learning

Abstract

The present invention relates to a method and apparatus for predicting tire wear using machine learning. The method for predicting tire wear in an apparatus mounted on a vehicle can include importing a tire wear database generated based on basic data, generating a dataset by preprocessing the basic data, classifying the dataset for each vehicle driving mode, optimizing hyperparameters for machine learning based on the classified dataset, and predicting a tire wear life of the vehicle by performing machine learning on the optimized hyperparameters.

Description

[0001] Cross-references to related applications

[0002] This application claims priority to Korean Patent Application No. 10-2020-0176787, filed on December 16, 2020, the entire contents of which are incorporated herein by reference for all purposes. Technical Field

[0003] This invention relates to a tire wear prediction technology, and more specifically, to a tire wear prediction technology utilizing artificial intelligence machine learning. Background Technology

[0004] Tire wear life is difficult to predict in terms of tire wear amount and tread pattern before actual testing / evaluation is performed.

[0005] Currently, tire wear performance is determined by actual road driving (or actual road simulation of indoor roller rotation test) after the tire sample is manufactured.

[0006] Tire finite element method (FEM) analysis predicts wear based on tire tread pattern, internal structure, and tread compound. However, due to its low correlation with actual driving conditions, its application in tire development and utilization is insufficient.

[0007] Accordingly, there is a need for a method that enables tire developers to quantitatively analyze tire wear and predict tire wear life in advance, taking proactive measures to address potential future problems.

[0008] The information disclosed in the background section of this invention is intended only to enhance the understanding of the overall background of the invention and should not be construed as an admission or in any way implying that such information constitutes prior art known to those skilled in the art. Summary of the Invention

[0009] Various aspects of the present invention are dedicated to providing a method and apparatus for predicting tire wear using artificial intelligence and machine learning.

[0010] Various aspects of the present invention provide a tire wear prediction method and apparatus using AI machine learning, the method being configured to facilitate the development of virtual vehicles by early prediction and analysis of tire wear performance when assessing tire wear before actual driving, thereby taking proactive measures to address future problems.

[0011] Various aspects of the present invention provide a multi-output regression prediction model that utilizes machine learning techniques, configured to quantitatively perform data-based tire wear prediction based on a learning / prediction database framework built for developing machine learning algorithms.

[0012] Various aspects of the present invention provide a tire wear prediction method and apparatus utilizing AI machine learning, the method being configured to facilitate real-time active control and operation management of autonomous vehicles / mobile vehicles by linking a multi-output regression prediction model to a vehicle controller, infotainment system, and mobile operator using artificial intelligence machine learning techniques.

[0013] Various aspects of the present invention provide a tire wear prediction method and apparatus using AI machine learning, the method being configured to perform relative comparison and contribution analysis of wear characteristics for each tire wear influencing factor.

[0014] The technical problems to be solved by the present invention are not limited to those described above. Those skilled in the art to which the various exemplary embodiments of the present invention pertain will clearly understand any other technical problems not mentioned herein through the following description.

[0015] According to various aspects of the present invention, a method for predicting tire wear in a device installed in a vehicle may include: importing a tire wear database generated based on basic data; generating a dataset by preprocessing the basic data; classifying the dataset for each vehicle driving mode; optimizing hyperparameters for machine learning based on the classified dataset; and predicting the tire wear life of the vehicle by performing machine learning on the optimized hyperparameters.

[0016] According to various exemplary embodiments of the present invention, the basic data may include explanatory and predictive variables obtained through tire wear tests on actual vehicles, wherein the explanatory variables include at least one of vehicle information, vehicle driving information, tire information and wheel alignment information, and the predictive variables include the wear life of each tire groove.

[0017] According to various exemplary embodiments of the present invention, preprocessing of basic data may include at least one of the following: converting discrete and qualitative variables into quantitative variables, normalizing quantitative variables, eliminating extreme values ​​from predictor variables, and compensating for missing values ​​of explanatory variables.

[0018] According to various exemplary embodiments of the present invention, hyperparameters can be optimized by eliminating factors with importance below a predetermined value from the categorized dataset through an importance analysis of each tire wear factor based on the Least Absolute Shrinkage and Selection Operator (LASSO) model.

[0019] According to various exemplary embodiments of the present invention, tire wear life is predicted based on multi-output regression analysis technology, and is predicted by multi-output regression analysis technology.

[0020] According to various exemplary embodiments of the present invention, multi-output regression analysis techniques may include random forest technique and stochastic gradient boosting technique, and one of random forest technique and stochastic gradient boosting technique may be selectively used based on the amount of data to be classified.

[0021] According to various exemplary embodiments of the present invention, the method may further include: outputting predicted tire life information via an output device disposed in the vehicle, and sending the predicted tire life information to other devices.

[0022] According to various exemplary embodiments of the present invention, other devices may include at least one of a vehicle controller, a vehicle developer server, a driver terminal, and a mobile operator server.

[0023] According to various exemplary embodiments of the present invention, the method may further include: optimizing vehicle driving-related parameters by performing active vehicle control based on predicted tire life information, and the active vehicle control may include at least one of braking control, suspension control, steering wheel control, and steering control.

[0024] According to various exemplary embodiments of the present invention, basic data can be collected from sensors installed in the vehicle, and the basic data may include at least one of the following: driving mode analysis information detected by an acceleration sensor built into the airbag control unit, wheel alignment change information detected by an electronic suspension device, vehicle weight change information detected by an automatic leveling device, tire pressure change information detected by a tire pressure monitoring system (TPMS), and driving climate environment information detected by an outside air temperature sensor in the air conditioning unit.

[0025] According to various aspects of the present invention, a tire wear prediction device may include a memory and a processor electrically connected to the memory, wherein the processor may import a tire wear database generated based on basic data from the memory or an external device, generate a dataset by preprocessing the basic data, classify the dataset for each vehicle driving mode, optimize hyperparameters for machine learning based on the classified dataset, and predict the tire wear life of the vehicle by performing machine learning on the optimized hyperparameters.

[0026] According to various exemplary embodiments of the present invention, the basic data may include explanatory and predictive variables obtained through tire wear tests of actual vehicles, and the explanatory variables may include at least one of vehicle information, vehicle driving information, tire information and wheel alignment information, and the predictive variables may include the wear life of each tire groove.

[0027] According to various exemplary embodiments of the present invention, the processor may include at least one of the following: means for converting discrete and qualitative variables into quantitative variables, means for normalizing quantitative variables, means for eliminating extreme values ​​in predictor variables, and means for compensating for missing values ​​of explanatory variables.

[0028] According to various exemplary embodiments of the present invention, hyperparameters can be optimized by eliminating factors with importance below a predetermined value from the categorized dataset, based on the Least Absolute Value Shrinkage and Selection Operator (LASSO) model, through importance analysis of each tire wear factor.

[0029] According to various exemplary embodiments of the present invention, tire wear life can be predicted based on multi-output regression analysis technology, and predicted by multi-output regression analysis technology.

[0030] According to various exemplary embodiments of the present invention, multi-output regression analysis techniques may include random forest techniques and stochastic gradient boosting techniques, and the processor may selectively select either random forest techniques or stochastic gradient boosting techniques based on the number of datasets to be classified.

[0031] According to various exemplary embodiments of the present invention, the processor may include at least one of the following: means for controlling the output of predicted tire life information or information processed based on the predicted tire life information via an output device disposed in a vehicle, and means for controlling the transmission of the predicted tire life information to other devices.

[0032] According to various exemplary embodiments of the present invention, other devices may include at least one of a vehicle controller, a vehicle developer server, a driver terminal, and a mobile operator server.

[0033] According to various exemplary embodiments of the present invention, the processor can optimize vehicle driving-related parameters by performing active vehicle control based on predicted tire life information, and the active vehicle control can include at least one of braking control, suspension control, steering wheel control, and steering control.

[0034] According to various exemplary embodiments of the present invention, basic data can be collected from sensors disposed in the vehicle, and the basic data may include at least one of the following: driving mode analysis information detected by an acceleration sensor built into the airbag control unit, wheel alignment change information detected by an electronic suspension device, vehicle weight change information detected by an automatic leveling device, tire pressure change information detected by a tire pressure monitoring system (TPMS), and driving climate environment information detected by an outside air temperature sensor in the air conditioning unit.

[0035] The technical problems to be solved by the various embodiments of the present invention are not limited to those described above, and any other technical problems not mentioned herein will be clearly understood by those skilled in the art to which the various exemplary embodiments of the present invention pertain from the following description.

[0036] The methods and apparatus of the present invention have other features and advantages that will be apparent from or will be set forth in detail in the accompanying drawings and following detailed description, which together serve to explain the particular principles of the invention. Attached Figure Description

[0037] Figure 1 This is a flowchart illustrating a method for predicting tire wear using machine learning according to various exemplary embodiments of the present invention;

[0038] Figure 2 This is a schematic diagram illustrating the basic data processing procedures for constructing a training dataset according to various exemplary embodiments of the present invention;

[0039] Figure 3 This is a schematic diagram illustrating the learning modeling process according to various exemplary embodiments of the present invention;

[0040] Figure 4 This is a schematic diagram used to describe learning algorithms according to various exemplary embodiments of the present invention;

[0041] Figure 5 This is a schematic diagram illustrating a data cleaning process for AI learning / prediction according to various exemplary embodiments of the present invention;

[0042] Figure 6 It shows according to Figure 5 A specific example of a data cleaning process used for AI learning / prediction;

[0043] Figure 7 This is a flowchart describing the operating mechanism of a tire wear prediction device according to various exemplary embodiments of the present invention;

[0044] Figure 8 This is a graph illustrating tire wear life based on actual vehicle driving assessment and tire wear life prediction results using a tire wear learning algorithm, according to various exemplary embodiments of the present invention; and

[0045] Figure 9 The results of AI learning predictions for the rear wheel wear life ratio based on changes in the rear wheel camber angle are shown.

[0046] It should be understood that the accompanying drawings are not drawn to scale, but rather are schematic simplifications illustrating various features to demonstrate the basic principles of the invention. Specific design features of the invention disclosed herein (including, for example, specific dimensions, orientations, positions, and shapes) will be determined in part by the specific environment in which they are intended for application and use.

[0047] Throughout these figures, multiple figures in the accompanying drawings refer to the same or equivalent parts of the invention, and the reference numerals in the drawings refer to the same or equivalent parts of the invention. Detailed Implementation

[0048] Various embodiments of the invention will now be described in detail with reference to examples of these embodiments, which are illustrated in the accompanying drawings and described below. Although the invention will be described in conjunction with exemplary embodiments thereof, it will be understood that this specification is not intended to limit the invention to those exemplary embodiments. On the contrary, the invention is intended to cover not only the exemplary embodiments thereof, but also various alternatives, modifications, equivalents, and other embodiments that may be included within the spirit and scope of the invention as defined by the appended claims.

[0049] In the following, various exemplary embodiments of the invention will be described in detail with reference to the exemplary accompanying drawings. When adding reference numerals to the components of each drawing, it should be noted that even when the same or equivalent components are shown in other drawings, they are represented by the same reference numerals. Furthermore, in describing exemplary embodiments of the invention, detailed descriptions of known features or functions will be omitted so as not to unnecessarily obscure the spirit of the invention.

[0050] In describing components according to various exemplary embodiments of the present invention, terms such as first, second, "A", "B", (a), (b), etc., may be used. These terms are intended only to distinguish one component from another, and they do not limit the nature, order, or sequence of the components. Unless otherwise defined, all terms (including technical or scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the various exemplary embodiments of the present invention pertain. Terms defined in general dictionaries should be interpreted as having the same meaning as in the context of the relevant technical field, and are not to be interpreted as having an idealized or overly formal meaning unless expressly defined as having such a meaning in this application.

[0051] In the following text, reference will be made to Figure 1 , Figure 2 , Figure 3 , Figure 4 , Figure 5 , Figure 6 , Figure 7 , Figure 8 as well as Figure 9 The embodiments of the present invention will be described in detail below.

[0052] Figure 1 This is a flowchart illustrating a method for predicting tire wear using machine learning according to various exemplary embodiments of the present invention.

[0053] according to Figure 1 The method of predicting tire wear life through basic data and AI modeling can be implemented on a computer system connected to vehicle electronic control systems, vehicle sensor systems, infotainment systems, mobile operator servers, user terminals, etc. In the following text, the device that performs the method of predicting tire wear using machine learning will be referred to as a tire wear prediction device.

[0054] Reference Figure 1 A tire wear database can be established in the tire wear prediction device based on basic data (S110). As various exemplary embodiments of the present invention, the tire wear database can be established on a separate server linked to the tire wear prediction device. Here, the basic data can be actual tire wear test data and can include vehicle information, vehicle driving characteristic information, tire specification information, vehicle condition information, etc. The basic data can be normalized and supplemented through data preprocessing.

[0055] The tire wear prediction device can determine a learning model based on predefined modeling rules (S120). The modeling rules can be configured to determine the expected wear life of each tire groove based on vehicle information, information about vehicle driving conditions, tire specification information, wheel alignment information, etc.

[0056] The tire wear prediction device can perform machine learning based on a multi-output regression method (S130). As examples, multi-output regression methods can utilize techniques such as random forests and stochastic gradient boosting. Random forests are configured to perform regression or classification analysis, but multi-output regression methods are not limited to these. The Least Absolute Shrinkage and Selection Operator (LASSO) technique can be used, which eliminates the coefficients of unimportant variables and determines variable importance through penalized regression.

[0057] Random forest is an ensemble learning method used for classification and regression analysis. For example, it can operate to output classification or average predictions from multiple decision trees built during training.

[0058] Stochastic gradient boosting is a regression analysis learning technique that generates predictive models in the form of decision tree ensembles, builds models in stages, and inductively generalizes models by enabling optimization of any differentiable loss function.

[0059] Depending on the size of the dataset, random forest or stochastic gradient boosting techniques can be used.

[0060] When the training dataset is small (e.g., around 1000 samples), a random forest model can be prioritized. As the dataset increases, both random forest and stochastic gradient boosting can be used simultaneously, with selective utilization based on performance. LASSO can be used to analyze the contribution of feature factors to tire wear performance.

[0061] After training and testing wear life on each tire groove and performing 5-fold cross-validation, the performance is validated by averaging the five performance metrics. Then, the optimal machine learning model can be utilized based on the size and characteristics of the dataset.

[0062] The LASSO model allows us to determine the degree of influence (coefficient) of variables by analyzing the importance of each wear factor. For example, variables with non-zero influence, as determined by the LASSO model, can be classified as important variables, while variables with zero influence can be classified as unimportant variables.

[0063] The tire wear prediction device can output the learning results through a set display (S140).

[0064] Figure 2This is a schematic diagram illustrating the basic data processing procedures for constructing a training dataset according to various exemplary embodiments of the present invention.

[0065] Reference Figure 2 Basic data can mainly include explanatory variables and predictor variables.

[0066] Explanatory variables can include vehicle information, vehicle driving information, tire information, wheel alignment information, etc.

[0067] Vehicle information may include at least one of the following: vehicle class information, vehicle model information, power source information, power source name information, displacement information, electric vehicle battery capacity information, transmission type information, output (PS) information, torque (Kgf.m) information, drive type (front / rear wheel drive) information, 2WD / 4WD information, suspension type (multi-link / CTBA) information, FR side weight (Kg) information, RR side weight (Kg) information, and front / rear tire pressure (psi) information.

[0068] Driving mode information may include at least one of the following: driving mode type (complex / high speed, etc.), wear measurement timing (initial / intermediate / end, etc.), and driving end date (date / season, etc.).

[0069] Tire information may include at least one of the following: manufacturer information, tire inner diameter (inches) information, tire tread pattern information and tread physical properties (Tg (°C)) information, front tire information and rear tire information.

[0070] Here, each front tire information and rear tire information may include at least one of the following: tread width (mm) information, aspect ratio (%) information, tread width / aspect ratio information, SW height, and initial groove depth (center / outer / shoulder) (mm) information.

[0071] Wheel alignment information can mainly include wheel alignment specification information and wheel alignment measurement load weight information.

[0072] Wear life information for each tire groove can include wear life information for the front / rear left tire and right tire.

[0073] Tire wear prediction devices can generate datasets by preprocessing basic data and normalizing the data. A tire wear database can include basic data and datasets generated based on that basic data.

[0074] Figure 3 This is a schematic diagram illustrating the learning modeling process according to various exemplary embodiments of the present invention.

[0075] like Figure 3As shown by reference numeral 310 in the attached figure, the minimum value among the wear life of the 24 grooves formed on the four tires of the front / rear and left / right wheels of the corresponding vehicle can be determined as the tire wear life of the corresponding vehicle.

[0076] Each tire may include two center grooves, shoulder grooves, and shoulder tread block grooves.

[0077] In this case, at a driving distance "K", the expected wear life "A" of each tire groove can be determined by the following equation.

[0078] A = K × (initial trench depth - 1.6) / (initial trench depth - final trench depth)

[0079] The initial groove depth is the groove depth when the travel distance is zero, and the final groove depth is the groove depth when the travel distance is K. In this case, the expected wear life of the vehicle's tire can be determined as the minimum of the expected wear life of all grooves of the tire.

[0080] like Figure 3 As shown by reference numeral 320 in the attached figure, during the learning modeling process, the variable factors used to predict tire wear life can include vehicle information, driving condition information, tire information, and wheel alignment information. In this case, the variable factors can be input into a predetermined modeling function f(x) that applies vehicle driving, tire tread, tire wear mechanisms, etc., so that the expected wear life of each tire groove can be determined.

[0081] Figure 4 This is a schematic diagram used to describe learning algorithms according to various exemplary embodiments of the present invention.

[0082] Since the wear life of the drive axle (driving force transmission) tires is usually low, the wear life of the tires of a corresponding vehicle can be determined by the wear life of the drive axle tires.

[0083] Therefore, when building AI prediction models, it may be more effective to classify the front-wheel drive vehicle and rear-wheel drive vehicle datasets for prediction, rather than using the entire actual test dataset.

[0084] For example, as shown by reference numeral 410, learning algorithms according to various exemplary embodiments of the present invention can be implemented to predict the wear life of the front tires or all tires when the drive mode is front-wheel drive. On the other hand, when the drive mode is rear-wheel drive, as shown by reference numeral 420, learning algorithms can be implemented to predict the wear life of the rear tires or all tires.

[0085] A 2×2 combination of prediction models can be configured for the drive mode (front-wheel drive / rear-wheel drive) and the mounting position of the tires to be predicted (drive axle / all tires).

[0086] Furthermore, the tire wear database, built based on basic data, can be categorized according to driving mode. Figure 4 As shown, it can generate separate prediction models.

[0087] Figure 5 This is a schematic diagram illustrating a data cleaning process for AI learning / prediction according to various exemplary embodiments of the present invention.

[0088] The tire wear prediction device can convert all discrete (categorical) variables / qualitative variables (e.g., power source type, tire manufacturer, and driving mode) into quantitative variables (S510). Figure 6 Figure 610 shows an example of the transformation of a quantitative variable used as a driving force for characteristic variables.

[0089] The tire wear prediction device can normalize quantitative variables (S520). For example, wheel alignment values ​​with small absolute values ​​can be normalized by replacing them with values ​​between 0 and 100. Figure 6 Figure 620 in the figure shows an example of normalization for the characteristic variable outer inclination angle.

[0090] The tire wear prediction device can supplement the predictor variable (S530). For example, it can eliminate extreme values ​​(outliers) in the predictor variable. For example, outliers in the target data can be eliminated based on 3 standard deviations (3 sigma). Here, 3 sigma can refer to values ​​ranging from the mean of the corresponding predicted value to three standard deviations on either side. Figure 6 Figure 630 shows an example of eliminating outliers in tire groove wear life as a predictor variable.

[0091] The tire wear prediction device can supplement explanatory variables (S540). For example, when missing values ​​exist in the previous dataset of explanatory variables, the corresponding explanatory variables can be supplemented by replacing the missing values ​​with the mean. For example, for tire tread physical property (Tg) values, after being divided into all seasons and summer patterns, the missing values ​​can be replaced with the mean within the corresponding category. Figure 6 Figure 640 shows an example of supplementing missing values ​​for the initial groove depth and tread physical properties of the front tire as explanatory variables.

[0092] The data can be refined through steps S510 to S540 above to construct the final training dataset (S550). Figure 6 Figure 650 shows an example of input variables (explanatory variables) / output variables (predictor variables) that have undergone data cleaning processes from S510 to S540.

[0093] Figure 7 This is a flowchart describing the operating mechanism of a tire wear prediction device according to various exemplary embodiments of the present invention.

[0094] Reference Figure 7 The tire wear prediction device can import a tire wear database generated based on basic data from internal memory (or an external device / server), and then preprocess the data stored in the tire wear database to generate a dataset (S710 to S720). Here, preprocessing may include transformation from discrete / qualitative variables to quantitative variables, normalization of quantitative variables, removal of extreme values ​​in predictor variables, and supplementation of explanatory variables with missing values.

[0095] The tire wear prediction device can classify the dataset for each vehicle drive mode (S730). Here, vehicle drive mode can include front-wheel drive, rear-wheel drive, and four-wheel drive (4WD).

[0096] The tire wear prediction device can optimize the hyperparameters (S740) to be applied to the learning algorithm.

[0097] The tire wear prediction device can perform machine learning based on optimized hyperparameters to predict tire wear performance (S750). In this case, the predicted tire wear performance results can be sent to an in-vehicle display device, vehicle controller, driver terminal, or vehicle manufacturer server, mobile operator server, etc., for various purposes.

[0098] For example, tire wear prediction results can be processed to generate information related to expected mileage and the expected rate of mileage change, and then this information can be provided to the driver through in-vehicle displays such as instrument clusters, head-up displays (HUDs), and the driver's smart devices.

[0099] As various exemplary embodiments of the present invention, tire wear prediction results can be used for active vehicle control based on tire wear conditions. For example, tire wear prediction results can be used to optimize vehicle driving-related parameters in real time, such as braking / cornering / suspension / steering / driving.

[0100] As another example, tire wear prediction results can be provided to mobile vehicle drivers and used for proactive tire maintenance strategies. For instance, tire wear prediction results can be processed to generate vehicle management-related information, such as information on tire position changes and replacement timing, as well as driving usage analysis for further use.

[0101] Tire wear prediction devices according to various exemplary embodiments of the present invention can obtain information from various sensors installed in a vehicle and perform machine learning for tire wear prediction.

[0102] For example, a tire wear prediction device can obtain information needed for driving pattern analysis, such as rapid acceleration, sudden braking, and steering, from an acceleration sensor built into the airbag control unit.

[0103] In addition, the tire wear prediction device can obtain vehicle wheel alignment information detected by the electronic suspension system, vehicle weight change information and vehicle attitude change information detected by the automatic leveling device, and tire pressure change information from the Tire Pressure Monitoring System (TPMS).

[0104] In addition, the tire wear prediction device can obtain information related to the driving climate environment from the outside temperature sensor built into the air conditioning system.

[0105] When learning using artificial intelligence algorithms, the resource usage (i.e., processing load) of the controller (CPU / processor) is high. However, with already trained artificial intelligence algorithms, the resource usage of the controller can be significantly reduced when processing / analyzing / judging sensor information.

[0106] Accordingly, the tire wear prediction device (or tire wear prediction processor) according to the exemplary embodiment can be mounted on and driven by at least one controller mounted on the vehicle. As an example, the tire wear prediction device can be mounted on and driven by a specific vehicle ECU or ACU.

[0107] Figure 8 This is a graph showing the tire wear life based on actual vehicle driving assessment and the predicted tire wear life using a tire wear learning algorithm according to various exemplary embodiments of the present invention.

[0108] Reference Figure 8 The tire wear life prediction results, based on the minimum tire wear life assessed from actual vehicle driving and the tire wear life prediction results learned by AI, show an error rate of approximately 5.9%.

[0109] Figure 9 The results of AI learning predictions for the rear wheel wear life ratio based on changes in the rear wheel camber angle are shown.

[0110] Reference Figure 9 When the AI ​​learning method according to various exemplary embodiments of the present invention is applied, the tire wear trend shown to vary according to the rear wheel camber angle is consistent.

[0111] Therefore, the operation of the methods or algorithms described in conjunction with the exemplary embodiments disclosed herein can be directly embodied in a hardware module, or a software module executed by a processor, or a combination of both. The software module can reside on a storage medium (i.e., memory and / or storage device), such as RAM, flash memory, ROM, EPROM, EEPROM, registers, hard disk, removable disk, and CD-ROM.

[0112] An exemplary storage medium can be coupled to a processor, which can read information from and write information to the storage medium. Alternatively, the storage medium can be integrated with the processor. The processor and storage medium can reside in an Application Specific Integrated Circuit (ASIC). The ASIC can reside in the user terminal. In another scenario, the processor and storage medium can exist as separate components of the user terminal.

[0113] The above description is merely an illustration of the technical concept of the present invention. Various modifications and variations can be made by those skilled in the art without departing from the essential characteristics of the invention. Accordingly, the exemplary embodiments included in the exemplary embodiments of the present invention are not intended to limit the technical concept of the invention, but are used to describe the invention, and the scope of the technical concept of the invention is not limited by the embodiments. The scope of protection of the present invention can be interpreted by the appended claims, and all technical concepts within the scope of the appended claims can be interpreted as being included within the scope of the present invention.

[0114] The advantage of this invention is that it provides a method and apparatus for predicting tire wear using artificial intelligence machine learning.

[0115] Furthermore, the present invention has the advantage of providing a tire wear prediction method and apparatus using AI machine learning, which is configured to facilitate the development of virtual vehicles by early prediction and analysis of tire wear performance when assessing tire wear before actual driving, thereby taking proactive measures to address future problems.

[0116] Furthermore, the present invention has the advantage of providing a multi-output regression prediction model that utilizes machine learning techniques, configured to quantitatively perform data-based tire wear prediction based on a learning / prediction database framework built for developing machine learning algorithms.

[0117] Furthermore, the present invention has the advantage of providing a tire wear prediction method and apparatus that utilizes AI machine learning. The method is configured to link multiple output regression prediction models to a vehicle controller, infotainment system, and mobile operator by utilizing artificial intelligence machine learning technology, thereby facilitating real-time active control and operation management of autonomous vehicles / mobile vehicles.

[0118] Furthermore, the present invention has the advantage of providing a tire wear prediction method and apparatus using AI machine learning, wherein the method is configured to perform relative comparison and contribution analysis of the wear characteristics of each tire wear influencing factor.

[0119] Furthermore, the present invention has the advantage of reducing costs by replacing actual vehicle tire wear assessment with wear prediction based on artificial intelligence machine learning.

[0120] Furthermore, various effects can be provided directly or indirectly through this invention.

[0121] For ease of interpretation and precise definition of the appended claims, the terms “above,” “below,” “inside,” “outside,” “up,” “down,” “upward,” “downward,” “front,” “back,” “inner side,” “outer side,” “inland,” “outer,” “inner,” “outer,” “front,” and “rear” are used to describe features of the exemplary embodiments with reference to the positions of these features shown in the accompanying drawings. It will be further understood that the term “connection” or its derivatives refer to both direct and indirect connections.

[0122] The foregoing description of specific exemplary embodiments of the invention is for illustrative and descriptive purposes. It is not intended to be exhaustive or to limit the invention to the precise forms disclosed, and it will be apparent that various modifications and variations are possible in light of the foregoing teachings. The exemplary embodiments were chosen and described to explain the specific principles of the invention and its practical application, thereby enabling others skilled in the art to utilize and implement various exemplary embodiments of the invention, as well as various alternatives and modifications thereof. The scope of the invention is intended to be defined by the appended claims and their equivalents.

Claims

1. A method for predicting tire wear in a device installed in a vehicle, the method comprising: The processor imports a tire wear database generated based on basic data. The processor creates the dataset by preprocessing the basic data; The processor classifies the dataset for each vehicle driving method; The processor optimizes the hyperparameters used for machine learning based on the classified dataset; The processor predicts the tire wear life of a vehicle by performing machine learning on information obtained from various sensors set up in the vehicle, based on optimized hyperparameters. The information obtained from various sensors installed in the vehicle includes wheel alignment information, vehicle weight change information, tire pressure change information, and information related to driving climate conditions. The method further includes: The processor performs active vehicle control based on predicted tire life information to optimize vehicle driving-related parameters. The active vehicle control includes at least one of braking control, suspension control, steering wheel control, and steering control.

2. The method for predicting tire wear in a device installed in a vehicle according to claim 1, wherein, The basic data includes explanatory and predictive variables obtained through tire wear tests on actual vehicles. The explanatory variables include at least one of vehicle information, vehicle driving information, tire information, and wheel alignment information, and the predictive variables include the wear life of each tire groove.

3. The method for predicting tire wear in a device installed in a vehicle according to claim 2, wherein, Preprocessing of basic data includes at least one of the following: Convert discrete and qualitative variables into quantitative variables; Normalize quantitative variables; Eliminate extreme values ​​from the predictor variables; Compensate for missing values ​​of explanatory variables.

4. The method for predicting tire wear in a device installed in a vehicle according to claim 1, wherein, Based on the minimum absolute value contraction and selection operator model, the hyperparameters are optimized by eliminating factors with importance below a predetermined value from the categorized dataset through importance analysis of each tire wear factor.

5. The method for predicting tire wear in a device installed in a vehicle according to claim 1, wherein, Tire wear life is predicted using multi-output regression analysis.

6. The method for predicting tire wear in a device installed in a vehicle according to claim 5, wherein, Multi-output regression analysis techniques include random forest and stochastic gradient boosting techniques. One of the techniques to selectively utilize is random forest or stochastic gradient boosting, depending on the number of datasets to be classified.

7. The method for predicting tire wear in a device installed in a vehicle according to claim 1, further comprising: The processor outputs predicted tire life information through an output device located in the vehicle; The processor sends the predicted tire life information to other devices, including at least one of a vehicle controller, a vehicle developer server, a driver terminal, and a mobile operator server.

8. An apparatus for predicting tire wear, the apparatus comprising: Memory; and The processor, which is electrically connected to the memory, The processor is configured to: import a tire wear database generated from basic data from a memory or external device; generate a dataset by preprocessing the basic data; classify the dataset for each vehicle drive mode; optimize hyperparameters for machine learning based on the classified datasets; and predict the tire wear life of the vehicle by performing machine learning on information obtained from various sensors installed in the vehicle based on the optimized hyperparameters. The information obtained from various sensors installed in the vehicle includes wheel alignment information, vehicle weight change information, tire pressure change information, and information related to driving climate conditions. The processor is configured as follows: Vehicle driving-related parameters are optimized by performing active vehicle control based on predicted tire life information. The active vehicle control includes at least one of braking control, suspension control, steering wheel control, and steering control.

9. The apparatus for predicting tire wear according to claim 8, wherein, The basic data includes explanatory and predictive variables obtained through tire wear tests on actual vehicles. The explanatory variables include at least one of vehicle information, vehicle driving information, tire information, and wheel alignment information, and the predictive variables include the wear life of each tire groove.

10. The apparatus for predicting tire wear according to claim 9, wherein, The processor is configured to process at least one of the following: Convert discrete and qualitative variables into quantitative variables; Normalize quantitative variables; Eliminate extreme values ​​of predictor variables; Compensate for missing values ​​of explanatory variables.

11. The apparatus for predicting tire wear according to claim 8, wherein, Based on the minimum absolute value contraction and selection operator model, the hyperparameters are optimized by eliminating factors with importance below a predetermined value from the categorized dataset through importance analysis of each tire wear factor.

12. The apparatus for predicting tire wear according to claim 8, wherein, Tire wear life is predicted using multi-output regression analysis.

13. The apparatus for predicting tire wear according to claim 12, wherein, Multi-output regression analysis techniques include random forest and stochastic gradient boosting techniques. The processor is configured to selectively choose between a random forest technique and a stochastic gradient boosting technique based on the number of datasets to be classified.

14. The apparatus for predicting tire wear according to claim 8, wherein, The processor is configured to process at least one of the following: controlling the output of predicted tire life information or information processed based on the predicted tire life information via an output device located in the vehicle; and controlling the transmission of the predicted tire life information to other devices, including at least one of a vehicle controller, a vehicle developer server, a driver terminal, and a mobile operator server.

Images