A real vehicle attribute prediction method and device, electronic equipment and storage medium

Abstract

The embodiment of the application provides a kind of real vehicle attribute prediction method, device, electronic equipment and storage medium, using preset tire design database and tire external characteristic database, corresponding first target prediction model is obtained by training, and using tire external characteristic database and real vehicle dynamics attribute database, corresponding second target prediction model is obtained by training, further, the high-precision tire external characteristic prediction result corresponding to the tire parameter of target tire is output using the first target prediction model trained, so that the second target prediction model tire is based on the tire external characteristic prediction result, further accurately predict the real vehicle attribute of vehicle level, using the above mode, effectively improve the prediction accuracy of real vehicle attribute, simultaneously, based on the prediction mode of the above machine learning, the prediction efficiency of real vehicle attribute is guaranteed, to further improve the production efficiency of vehicle development.

Description

Technical Field

[0001] The present invention relates to the technical field of real vehicle dynamics property prediction, and in particular to a real vehicle property prediction method, device, electronic equipment and storage medium. Background Art

[0002] With the gradual advancement of intelligent driving and the Internet of Things (IoE), the prediction of vehicle properties (such as load) based on real-vehicle dynamics prediction models has become increasingly prominent in vehicle development, design, and practical planning. Tires, as key components that come into direct contact with the ground during driving, are crucial for dynamic control and trajectory tracking. Therefore, there is an urgent need for a high-precision real-vehicle property prediction method based on tire dynamics to accurately predict the actual stress conditions of vehicle tires during driving. This is crucial for improving the prediction accuracy of real-vehicle properties and the production efficiency of vehicle development. Summary of the Invention

[0003] The embodiments of the present application provide a method and related apparatus for predicting the properties of a real vehicle based on tire dynamic characteristics, which are used to improve the prediction accuracy of the properties of the real vehicle.

[0004] In a first aspect, an embodiment of the present application provides a method for predicting attributes of a real vehicle, including:

[0005] A tire design database and a tire external characteristic database are used to train a preset first prediction model to obtain a corresponding first target prediction model.

[0006] The tire external characteristic database and the real vehicle dynamic property database are used to train the preset second prediction model to obtain a corresponding second target prediction model.

[0007] Inputting tire parameters of a target tire into the trained first target prediction model to obtain a tire external characteristic prediction result output by the first target prediction model based on the tire parameters;

[0008] The tire external characteristic prediction result is input into the trained second target prediction model to obtain the actual vehicle attribute prediction result output by the second target prediction model.

[0009] In a second aspect, an embodiment of the present application provides a real vehicle attribute prediction device, comprising:

[0010] The first training module is used to train a preset first prediction model using a tire design database and a tire external characteristic database to obtain a corresponding first target prediction model.

[0011] The second training module is used to train a preset second prediction model using the tire external characteristic database and the real vehicle dynamic attribute database to obtain a corresponding second target prediction model.

[0012] an external characteristic prediction module, configured to input tire parameters of a target tire into the trained first target prediction model, and obtain a tire external characteristic prediction result output by the first target prediction model based on the tire parameters;

[0013] The attribute prediction module is used to input the tire external characteristic prediction result into the trained second target prediction model to obtain the actual vehicle attribute prediction result output by the second target prediction model.

[0014] In an optional embodiment, before the tire design database and the tire external characteristic database are used to train the preset first prediction model, the first training module is further configured to:

[0015] Cleaning rules are used to clean the obtained tire design data and construct a corresponding tire design database, wherein the tire design data includes design parameters of at least one vehicle tire and parameter ranges determined for the design parameters of the at least one vehicle tire.

[0016] The cleaning rule is adopted to perform data cleaning on the obtained tire external characteristic data, and a corresponding tire external characteristic database is constructed, wherein the tire external characteristic data includes at least one external characteristic parameter of a vehicle tire.

[0017] The cleaning rules are used to clean the acquired real vehicle dynamics attribute data and construct a corresponding real vehicle dynamics attribute database, wherein the real vehicle dynamics attribute database includes the real vehicle dynamics attributes of at least one vehicle and attribute ranges determined for the real vehicle dynamics attributes of the at least one vehicle.

[0018] In an optional embodiment, the tire design database and the tire external characteristic database are used to train a preset first prediction model to obtain a corresponding first target prediction model, and the first training module is specifically used to:

[0019] A tire design database and a tire external characteristic database are used to set a corresponding first training sample set, wherein a first training sample includes: first input information determined based on the tire design database and first standard information determined based on the tire external characteristic database.

[0020] Using the first training sample in the first training sample set, the preset first prediction model is subjected to multiple rounds of iterative training, and when the preset first convergence condition is met, the first target prediction model is output; wherein, during one round of iterative training, the following operations are performed: using the first prediction model, based on the first input information in the first training sample, the corresponding external characteristic prediction result is obtained, and based on the loss value between the external characteristic prediction result and its corresponding first standard information, the parameters of the first prediction model are adjusted.

[0021] In an optional embodiment, after obtaining the corresponding first target prediction model, the first training module is further used to:

[0022] Using a preset first learning rate, multiple rounds of hyperparameter tuning are performed on the obtained first target prediction model, wherein, during one round of hyperparameter tuning, the following operations are performed: based on the cross-validation value of the first target prediction model, the hyperparameters of the first target prediction model are adjusted, and when the adjusted first target prediction model meets the preset first verification condition, the first learning rate of the first target prediction model is reduced.

[0023] In an optional embodiment, the tire external characteristics database and the real vehicle dynamics attribute database are used to train a preset second prediction model to obtain a corresponding second target prediction model, and the second training module is specifically configured to:

[0024] The tire external characteristic database and the real vehicle dynamic property database are used to set a corresponding second training sample set, wherein a second training sample includes: second input information determined based on the tire external characteristic database and second standard information determined based on the real vehicle dynamic property database.

[0025] Using the second training sample in the second training sample set, the preset second prediction model is subjected to multiple rounds of iterative training, and when the preset second convergence condition is met, the second target prediction model is output; wherein, during one round of iterative training, the following operations are performed: using the second prediction model, based on the second input information in the second training sample, the corresponding real vehicle attribute prediction result is obtained, and based on the loss value between the real vehicle attribute prediction result and its corresponding second standard information, the parameters of the second prediction model are adjusted.

[0026] In an optional embodiment, after obtaining the corresponding second target prediction model, the second training module is further used to:

[0027] Using a preset second learning rate, multiple rounds of hyperparameter tuning are performed on the obtained second target prediction model, wherein, during one round of hyperparameter tuning, the following operations are performed: based on the cross-validation value of the second target prediction model, the hyperparameters of the second target prediction model are adjusted, and when the adjusted second target prediction model meets the preset second verification condition, the second learning rate of the second target prediction model is reduced.

[0028] In a third aspect, an electronic device is proposed, comprising a processor and a memory, wherein the memory stores program code, and when the program code is executed by the processor, the processor executes the steps of the real vehicle attribute prediction method described in the first aspect above.

[0029] In a fourth aspect, a computer-readable storage medium is proposed, which includes a program code. When the program code is run on an electronic device, the program code is used to enable the electronic device to execute the steps of the real vehicle attribute prediction method described in the first aspect.

[0030] In a fifth aspect, a computer program product is proposed, which is used to execute the steps of the real vehicle attribute prediction method described in the first aspect.

[0031] The technical effects of the embodiments of this application are as follows:

[0032] The embodiments of the present application provide a method, device, electronic device and storage medium for predicting the properties of a real vehicle. A preset tire design database and a tire external characteristic database are used to train a corresponding first target prediction model, and a tire external characteristic database and a real vehicle dynamic property database are used to train a corresponding second target prediction model. The tire design database and the tire external characteristic database respectively record the design parameters and dynamic external characteristic parameters of the vehicle tire under real conditions. Based on the above method, the first target prediction model can output a high-precision tire external characteristic prediction result based on the input tire parameters, and enable the second target prediction model tire to further accurately predict the real vehicle properties at the vehicle level based on the tire external characteristic prediction result, thereby effectively improving the prediction accuracy of the real vehicle properties. At the same time, the prediction method based on the above machine learning ensures the prediction efficiency of the real vehicle properties, thereby further improving the production efficiency of vehicle development. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 A schematic diagram of a possible application scenario provided by an embodiment of the present application;

[0034] Figure 2 A flowchart of a method for predicting vehicle attributes provided in an embodiment of the present application;

[0035] Figure 3 A schematic diagram of a model training provided in an embodiment of the present application;

[0036] Figure 4 This is a flowchart illustrating a method for predicting vehicle attributes according to an embodiment of the present application;

[0037] Figure 5 A logic diagram illustrating a method for predicting vehicle attributes according to an embodiment of the present application;

[0038] Figure 6 A schematic diagram of the structure of a real vehicle attribute prediction device provided in an embodiment of the present application;

[0039] Figure 7 A schematic diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0040] The following will be combined with the accompanying drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the embodiments described are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0041] It should be noted that in the description of this application, "multiple" is understood to mean "at least two." "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. A and B are connected, which can mean: A and B are directly connected, and A and B are connected through C. In addition, in the description of this application, words such as "first" and "second" are only used for the purpose of distinguishing descriptions, and cannot be understood as indicating or implying relative importance, nor can they be understood as indicating or implying an order.

[0042] In addition, the collection, dissemination, and use of data in the technical solution of this application comply with the requirements of relevant national laws and regulations.

[0043] The design ideas of the embodiments of this application are as follows:

[0044] With the gradual upgrading of intelligent driving and the Internet of Things, the prediction of real vehicle properties (such as load, etc.) based on real vehicle dynamics prediction models has become increasingly prominent in vehicle development, design and actual planning. How to further improve the prediction accuracy of real vehicle properties has become an important issue that needs to be solved urgently.

[0045] In order to improve the prediction accuracy of real vehicle properties, the embodiments of the present application provide a real vehicle property prediction method, device, electronic device and storage medium, which uses a preset tire design database and a tire external characteristic database to train and obtain a corresponding first target prediction model, and uses a tire external characteristic database and a real vehicle dynamic property database to train and obtain a corresponding second target prediction model, wherein the tire design database and the tire external characteristic database respectively record the design parameters and dynamic external characteristic parameters of the vehicle tire under real conditions. Based on the above method, the first target prediction model can output a high-precision tire external characteristic prediction result based on the input tire parameters, and enable the second target prediction model tire to further accurately predict the real vehicle properties at the vehicle level based on the tire external characteristic prediction result, thereby effectively improving the prediction accuracy of the real vehicle properties. At the same time, based on the above-mentioned machine learning prediction method, the prediction efficiency of the real vehicle properties is guaranteed, thereby further improving the production efficiency of vehicle development.

[0046] Furthermore, based on the above design ideas, the real vehicle attribute prediction method provided in the embodiment of the present application can be executed by an internal computer program in one or more storage media associated with the user terminal, or it can be executed by an external electronic device that has a communication connection with the user terminal and returns the corresponding real vehicle attribute prediction results to the user terminal. This application does not impose any restrictions on this.

[0047] See Figure 1 FIG2 is a schematic diagram of a possible application scenario provided by an embodiment of the present application, which includes a target terminal 101 and an optional server 102. The target terminal 101 is a terminal held by a user, and information can be exchanged between the target terminal 101 and the server 102 via a communication network. The communication mode adopted by the communication network may include wireless communication mode and wired communication mode.

[0048] Exemplarily, the target terminal 101 can access the network through cellular mobile communication technology and communicate with the server 102, and the cellular mobile communication technology includes the fifth generation mobile communication (5th Generation Mobile Networks, 5G) technology; optionally, the target terminal 101 can also access the network through short-range wireless communication and communicate with the target cloud server 102, and the short-range wireless communication method includes Wireless Fidelity (Wi-Fi) technology.

[0049] The present application embodiment does not impose any restrictions on the number of the above devices. Figure 1 As shown, only the target terminal 101 and the server 102 are described as examples, and the above devices and their respective functions are briefly introduced below.

[0050] The target terminal 101 is a device that can provide voice and / or data connectivity to users, including: a handheld terminal device with a wireless connection function, a vehicle-mounted terminal device, etc.

[0051] Exemplarily, the target terminal 101 includes but is not limited to: Android devices, IOS devices, mobile phones, tablet computers, laptops, PDAs, mobile Internet devices (MID), wearable devices, virtual reality (VR) devices, augmented reality (AR) devices, wireless terminal devices in industrial control, wireless terminal devices in unmanned driving, wireless terminal devices in smart grids, wireless terminal devices in transportation safety, wireless terminal devices in smart cities, or wireless terminal devices in smart homes, etc.

[0052] For example, in an embodiment of the present application, a user can send a user request related to the prediction of real vehicle attributes to the server 102 through the running client (such as an APP, browser, short video software, or a web page, mini-program, etc.) in the target terminal 101, so that the server 102 responds to the user request and predicts the real vehicle attributes and returns the corresponding prediction results through its internally trained first target prediction model and second target prediction model.

[0053] It will be understood that the above method is only an example. In an optional embodiment, the user can also obtain the prediction results of the real vehicle attributes through the first target prediction model and the second target prediction model sealed in the storage medium associated with the target terminal 101. This application will not go into details here.

[0054] Furthermore, server 102 can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers. It can also be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms, which will not be elaborated here.

[0055] Based on the above application scenarios, the real vehicle attribute prediction method provided by the embodiment of the present application will be further described and illustrated with reference to the accompanying drawings. Figure 2 Shown, including:

[0056] S201: Using a tire design database and a tire external characteristic database, a preset first prediction model is trained to obtain a corresponding first target prediction model.

[0057] Specifically, the embodiments of the present application can construct corresponding tire design databases and tire external characteristic databases based on the tire design data and tire external characteristic data collected under actual conditions, wherein the above-mentioned tire design data and tire external characteristic data can be directly obtained from the tire design scheme under actual conditions, or can be obtained based on relevant technical simulations, and the present application does not impose any restrictions on this.

[0058] For example, referring to Table 1 below, the tire design data may at least involve the following:

[0059] Table 1

[0060] Involvement direction of design parameters Number Hysteresis loss DMA of rubber formulation ≈50 Raw material, strength, density, damping, etc. ≈30 Parameters related to tire section and inner cavity design ≈30 Related parameters of tire pattern design ≈30

[0061] For example, referring to Table 2 below, the tire external characteristic data may at least include the following:

[0062] Table 2

[0063]

[0064] Furthermore, in the embodiment of the present application, cleaning rules are used to clean the obtained tire design data and tire external characteristic data, respectively, and to construct corresponding tire design databases and tire external characteristic databases. Specifically, the cleaning rules may include:

[0065] 1) Data supplement.

[0066] 2) Data analogy.

[0067] 3) The data are deemed to be the same.

[0068] 4) Data correction.

[0069] Specifically, if the tire design data and / or tire external characteristic data contain some parameter data that are not clearly represented, such data can be supplemented, corrected, or deemed equivalent by humans or machines to clarify the actual significance of the data representation; optionally, in order to improve the efficiency of obtaining tire design data and / or tire external characteristic data, data analogy can also be performed on some iconic parameter data in the tire design data and / or tire external characteristic data to obtain richer training data required by the model.

[0070] For example, assuming that the tire external characteristic data contains the parameter data "resistance: 100N" which is not clearly represented, the above-mentioned cleaning rules are used to correct it to the data "rolling resistance: 100N" that is consistent with the external characteristic parameters shown in Table 2 above, so as to clarify the actual meaning of the parameter data.

[0071] Furthermore, based on the cleaned tire design data, parameter ranges corresponding to the tire design parameters are defined through physics, chemistry, acoustics, etc., and a corresponding tire design database is constructed.

[0072] For example, physical laws (e.g., Newton's second law of motion F=m*a; density formula m=ρ*V; ideal gas law P*V=n*R*T; acceleration formula a=v / t, etc.) are used to define a parameter range corresponding to the design parameters of the tire. The parameter range may include: force, mass, density, volume, pressure, molar weight of air molecules, temperature, speed, time, etc., so that the model can learn tire-related design parameters.

[0073] For another example, in a more specific embodiment, corresponding parameter ranges can be set in the tire design parameters corresponding to the actual requirements of the tire external characteristic parameters to be learned by the model. For example, when the tire external characteristic parameters to be learned are defined as: tire rolling grip, the parameter ranges related to the design parameters that can be set are: {tire weight, lateral stiffness, tire groove depth, rubber viscoelasticity, geometric cross-sectional dimensions, vertical dynamic stiffness, longitudinal slip coefficient, dynamic load radius}.

[0074] Furthermore, in an optional embodiment, in order to reduce the problem of inaccurate collection of some tire design data and / or tire external characteristic data under actual conditions, the technical experience of technical personnel in this field under actual measurements can be used to supplement the above-mentioned collected tire design data and / or tire external characteristic data through relevant vehicle simulation models, so as to ensure the accuracy of the sample data used in model training.

[0075] For example, the CAE simulation model CDTIRE is used to perform vehicle simulation on the actually measured tire design data, and the corresponding simulated tire data can also be regarded as the above-mentioned collected tire external characteristic data, which can be added to the constructed tire external characteristic database after being cleaned by the above-mentioned data.

[0076] It is understandable that the above method is only an example. In actual situations, the above tire design data and / or tire external characteristic data can be supplemented and adjusted through the technical experience of technical personnel in this field, or some other vehicle simulation models, such as FTIRE, PAC, etc., can be used to perform vehicle simulation based on the actual measured tire design data. This application will not go into details here.

[0077] Furthermore, in an optional embodiment, based on the tire design database and tire external characteristic database constructed above, a preset first prediction model is trained to obtain a corresponding first target prediction model, specifically comprising the following steps:

[0078] Step 1: Using a tire design database and a tire external characteristic database, set a corresponding first training sample set, wherein a first training sample includes: first input information determined based on the tire design database and first standard information determined based on the tire external characteristic database.

[0079] Step 2: Use the first training sample in the first training sample set to perform multiple rounds of iterative training on the preset first prediction model, and output the first target prediction model when the preset first convergence condition is met; wherein, during one round of iterative training, the following operations are performed: use the first prediction model to obtain the corresponding external characteristic prediction result based on the first input information in the first training sample, and adjust the parameters of the first prediction model based on the loss value between the external characteristic prediction result and its corresponding first standard information.

[0080] Specifically, a machine learning algorithm is used to train the first prediction model based on the tire design database and tire external characteristic database constructed above, and the parameters of the model are iteratively adjusted according to the loss function and / or cost function corresponding to the algorithm until a first target prediction model that meets the preset first convergence condition is obtained.

[0081] Exemplarily, the embodiment of the present application uses a machine learning algorithm such as BOOST to perform multiple rounds of iterative training on the first prediction model. In each iterative training process, the loss value between the external characteristic prediction result output by the model and the first standard information corresponding to the training is calculated by using the loss function and / or cost function corresponding to the algorithm, and based on the loss value, the parameters of the first prediction model are adjusted. The loss function is as follows:

[0082]

[0083] Among them, m is the number of first training samples, h θ (x (i) ) is the external feature prediction result corresponding to the first training sample i, y (i) is the first standard information corresponding to the i-th first training sample.

[0084] Optionally, a multi-threaded algorithm is used for model training, such as ADABOOST, XGBOOST, RANDOM, FOREST, KNN and other machine learning related algorithms. Different algorithms are evaluated and the optimal model is selected as the first target prediction model.

[0085] Furthermore, in an optional embodiment, based on the above method, after obtaining the corresponding first target prediction model, the method further includes:

[0086] Step 3: Use a preset first learning rate to perform multiple rounds of hyperparameter tuning on the obtained first target prediction model; wherein, during one round of hyperparameter tuning, the following operations are performed: based on the cross-validation value of the first target prediction model, the hyperparameters of the first target prediction model are adjusted, and when the adjusted first target prediction model meets the preset first verification condition, the first learning rate of the first target prediction model is reduced.

[0087] Specifically, in order to further improve the prediction performance of the model, the set first learning rate is used to perform multiple rounds of hyperparameter tuning on the trained first target prediction model, wherein each first parameter estimator is used to independently predict the external characteristic parameters of the tire. The hyperparameters may include (taking XGBOOST as an example): the number of parameter estimators n_estimators, the maximum depth of the decision tree max_depth, the deepest depth of the decision tree min_weight, the gamma parameter gamma, etc., and the adjustment order of each hyperparameter can be manually specified by those skilled in the art, which will not be repeated here.

[0088] Exemplarily, in the machine learning algorithm XGBOOST, a set first learning rate (represented here by the estimated time eta) is used to perform multiple rounds of hyperparameter tuning on the trained first target prediction model. Taking the first tuning process as an example, based on the above-mentioned first learning rate, the gamma parameter gamma, the maximum depth of the decision tree max_depth, the deepest depth of the decision tree min_weight, the minimum sub-case sum min_child_weight, the subsample training ratio subsample, the feature sampling ratio colsample_bytree and the regularization parameters (alpha, lambda) of the first target prediction model are tuned in turn. The first tuning result is: {eta=0.01; gamma=0.05; max_depth=3; min_child_weight=1; subsample=0.6; colsample_bytree=0.6; alpha=0;}. When it is determined that the adjusted first target prediction model meets the preset first verification condition (such as, the score reaches the preset threshold), the first learning rate of the first target prediction model is reduced until the model is optimal.

[0089] S202: Using the tire external characteristic database and the actual vehicle dynamic property database, a preset second prediction model is trained to obtain a corresponding second target prediction model.

[0090] Specifically, the embodiments of the present application can construct a corresponding real vehicle dynamics attribute database based on the real vehicle dynamics attribute data under actual conditions, wherein the above-mentioned real vehicle dynamics attribute data can be directly obtained from the vehicle dynamics data measured under real conditions, or can be obtained based on relevant technical simulations (such as a real vehicle dynamics test bench), and this application does not impose any restrictions on this.

[0091] For example, referring to Table 2 below, the actual vehicle dynamics attribute data may at least involve the following:

[0092] Table 3

[0093] Real vehicle dynamics properties Vehicle dynamic handling stability Vehicle comfort Vehicle quietness Dynamic model chassis algorithm model intervention Chassis integrated algorithm model intervention Vehicle safety on low-adhesion roads and electronic control systems

[0094] Furthermore, in an embodiment of the present application, cleaning rules are used to clean the obtained real vehicle dynamics attribute data and construct a corresponding real vehicle dynamics attribute database. Specifically, the cleaning rules can be the same as the cleaning rules in the above S201 and will not be repeated here.

[0095] Furthermore, based on the cleaned real vehicle dynamics attribute data, the attribute range of the real vehicle dynamics attribute of at least one vehicle is defined according to the whole vehicle attribute requirements, and a corresponding real vehicle dynamics attribute database is constructed.

[0096] Exemplarily, based on the requirements for vehicle attributes, the attribute ranges of the actual vehicle dynamic attributes of at least one vehicle are defined to include: weight, volume, etc.; optionally, those skilled in the art may also flexibly set or supplement the above attribute ranges according to the business needs under actual conditions, and this application does not impose any restrictions on this.

[0097] Furthermore, in an optional embodiment, based on the above-constructed real vehicle dynamics attribute database, a preset second prediction model is trained to obtain a corresponding second target prediction model, which specifically includes the following steps:

[0098] Step 1: Using the tire external characteristic database and the actual vehicle dynamic property database, set a corresponding second training sample set, wherein a second training sample includes: second input information determined based on the tire external characteristic database and second standard information determined based on the actual vehicle dynamic property database.

[0099] Step 2: Use the second training sample in the second training sample set to perform multiple rounds of iterative training on the preset second prediction model, and output the second target prediction model when the preset second convergence condition is met; wherein, during one round of iterative training, the following operations are performed: use the second prediction model to obtain the corresponding real vehicle attribute prediction result based on the second input information in the second training sample, and adjust the parameters of the second prediction model based on the loss value between the real vehicle attribute prediction result and its corresponding second standard information.

[0100] Specifically, a machine learning algorithm is used to train the second prediction model based on the constructed tire external characteristics database and the actual vehicle dynamics property database, and the parameters of the model are iteratively adjusted according to the loss function and / or cost function corresponding to the algorithm until a second target prediction model that meets the preset second convergence condition is obtained.

[0101] Exemplarily, a machine learning algorithm such as BOOST is used to perform multiple rounds of iterative training on the second prediction model. Specifically, during each iterative training process, the loss function and / or cost function corresponding to the algorithm is used to calculate the loss value between the actual vehicle attribute prediction result output by the model and the second standard information corresponding to the training, and the parameters of the second prediction model are adjusted accordingly.

[0102] It is understandable that in the above training process, multi-threaded algorithms are also used for model training. For example, ADABOOST, XGBOOST, RANDOM, FOREST, KNN and other machine learning related algorithms can be used for multi-threaded model training, and different algorithms can be evaluated to select the optimal model as the above-mentioned second target prediction model.

[0103] Furthermore, in an optional embodiment, based on the above method, after obtaining the corresponding second target prediction model, the method further includes:

[0104] Step 3: Use a preset second learning rate to perform multiple rounds of hyperparameter tuning on the obtained second target prediction model, wherein, during one round of hyperparameter tuning, the following operations are performed: based on the cross-validation value of the second target prediction model, the hyperparameters of the second target prediction model are adjusted, and when the adjusted second target prediction model meets the preset second verification condition, the second learning rate of the second target prediction model is reduced.

[0105] Exemplarily, in order to further improve the prediction performance of the model, the set second learning rate is used to perform multiple rounds of hyperparameter tuning on the trained second target prediction model. Taking the first tuning process as an example, based on the above-mentioned second learning rate, the number of parameter estimators n_estimators, the maximum depth of the decision tree max_depth, the deepest depth of the decision tree min_weight, the gamma parameter gamma, the subsample training ratio subsample, the feature sampling ratio colsample_bytree and the regularization parameter of the second target prediction model are tuned in turn. When the adjusted second target prediction model meets the preset second verification condition (such as, the score reaches the preset threshold), the first learning rate of the first target prediction model is reduced.

[0106] Based on the above method, the first target prediction model and the second target prediction model related to the external characteristics of the tire and the actual vehicle dynamic properties are respectively trained and obtained. The obtained first target prediction model and the second target prediction model can be sealed in a designated storage medium or uploaded to a cloud environment for the convenience of users' reasonable use.

[0107] S203: Inputting tire parameters of the target tire into the trained first target prediction model to obtain tire external characteristic prediction results output by the first target prediction model based on the tire parameters.

[0108] S204: Inputting the tire external characteristic prediction result into the trained second target prediction model to obtain the actual vehicle attribute prediction result output by the second target prediction model.

[0109] For further information, see Figure 3 As shown, based on the above method, the corresponding first target prediction model and the second target prediction model are trained and obtained, and the first target prediction model and the second target prediction model can also supplement the above-constructed tire external characteristic database and the actual vehicle dynamic property database through the predicted tire external characteristic data and the actual vehicle property data.

[0110] See Figure 4As shown, a flow chart of a real vehicle attribute prediction method provided in an embodiment of the present application is shown. After the collected tire design data, tire external characteristic data, and real vehicle attribute data are cleaned respectively, corresponding tire design databases, tire external characteristic databases, and real vehicle dynamic attribute databases are constructed based on the set attribute range and parameter range, and the preset first prediction model and second prediction model are trained accordingly. During the training process, the model parameters are iteratively adjusted through the loss / evaluation function, and multi-threaded evaluation is performed based on the preset multiple machine learning algorithms until the required first target prediction model and second target prediction model are obtained. Furthermore, the first target prediction model and the second target prediction model are hyperparameter tuned respectively, and the optimal first target prediction model and the optimal second target prediction model finally determined can be sealed in a designated storage medium according to actual business needs.

[0111] See Figure 5 As shown, a logic example diagram of a real vehicle attribute prediction method provided in an embodiment of the present application is provided. Through the specified tire design data source, tire external characteristic data source, and real vehicle attribute data source, corresponding tire design database, tire external characteristic database, and real vehicle attribute database are respectively constructed, and the above databases are used to implement the training of the first target prediction model and the second target prediction model. By uploading the above databases and models to the specified cloud environment, users can realize the external characteristic attribute prediction of the specified target tire and the real vehicle attribute prediction. This method effectively improves the prediction accuracy of the real vehicle attributes and ensures the prediction efficiency of the real vehicle attributes.

[0112] Furthermore, based on the same technical concept, the embodiment of the present application also provides a real vehicle attribute prediction device, which is used to implement the above method flow of the embodiment of the present application. Figure 6 As shown, the apparatus includes: a first training module 601, a second training module 602, an external characteristic prediction module 603 and an attribute prediction module 604, wherein:

[0113] The first training module 601 is used to train a preset first prediction model using a tire design database and a tire external characteristic database to obtain a corresponding first target prediction model.

[0114] The second training module 602 is used to train a preset second prediction model using the tire external characteristic database and the real vehicle dynamic attribute database to obtain a corresponding second target prediction model.

[0115] an external characteristic prediction module 603, configured to input tire parameters of a target tire into the trained first target prediction model, and obtain a tire external characteristic prediction result output by the first target prediction model based on the tire parameters;

[0116] The attribute prediction module 604 is configured to input the tire external characteristic prediction result into the trained second target prediction model to obtain the actual vehicle attribute prediction result output by the second target prediction model.

[0117] In an optional embodiment, before the tire design database and the tire external characteristic database are used to train the preset first prediction model, the first training module 601 is further configured to:

[0118] Cleaning rules are used to clean the obtained tire design data and construct a corresponding tire design database, wherein the tire design data includes design parameters of at least one vehicle tire and parameter ranges determined for the design parameters of the at least one vehicle tire.

[0119] The cleaning rule is adopted to perform data cleaning on the obtained tire external characteristic data, and a corresponding tire external characteristic database is constructed, wherein the tire external characteristic data includes at least one external characteristic parameter of a vehicle tire.

[0120] The cleaning rules are used to clean the acquired real vehicle dynamics attribute data and construct a corresponding real vehicle dynamics attribute database, wherein the real vehicle dynamics attribute database includes the real vehicle dynamics attributes of at least one vehicle and attribute ranges determined for the real vehicle dynamics attributes of the at least one vehicle.

[0121] In an optional embodiment, the tire design database and the tire external characteristic database are used to train a preset first prediction model to obtain a corresponding first target prediction model. The first training module 601 is specifically configured to:

[0122] A tire design database and a tire external characteristic database are used to set a corresponding first training sample set, wherein a first training sample includes: first input information determined based on the tire design database and first standard information determined based on the tire external characteristic database.

[0123] Using the first training sample in the first training sample set, the preset first prediction model is subjected to multiple rounds of iterative training, and when the preset first convergence condition is met, the first target prediction model is output; wherein, during one round of iterative training, the following operations are performed: using the first prediction model, based on the first input information in the first training sample, the corresponding external characteristic prediction result is obtained, and based on the loss value between the external characteristic prediction result and its corresponding first standard information, the parameters of the first prediction model are adjusted.

[0124] In an optional embodiment, after obtaining the corresponding first target prediction model, the first training module 601 is further configured to:

[0125] Using a preset first learning rate, multiple rounds of hyperparameter tuning are performed on the obtained first target prediction model, wherein, during one round of hyperparameter tuning, the following operations are performed: based on the cross-validation value of the first target prediction model, the hyperparameters of the first target prediction model are adjusted, and when the adjusted first target prediction model meets the preset first verification condition, the first learning rate of the first target prediction model is reduced.

[0126] In an optional embodiment, the tire external characteristic database and the real vehicle dynamic property database are used to train a preset second prediction model to obtain a corresponding second target prediction model. The second training module 602 is specifically configured to:

[0127] The tire external characteristic database and the real vehicle dynamic property database are used to set a corresponding second training sample set, wherein a second training sample includes: second input information determined based on the tire external characteristic database and second standard information determined based on the real vehicle dynamic property database.

[0128] Using the second training sample in the second training sample set, the preset second prediction model is subjected to multiple rounds of iterative training, and when the preset second convergence condition is met, the second target prediction model is output; wherein, during one round of iterative training, the following operations are performed: using the second prediction model, based on the second input information in the second training sample, the corresponding real vehicle attribute prediction result is obtained, and based on the loss value between the real vehicle attribute prediction result and its corresponding second standard information, the parameters of the second prediction model are adjusted.

[0129] In an optional embodiment, after obtaining the corresponding second target prediction model, the second training module 602 is further configured to:

[0130] Using a preset second learning rate, multiple rounds of hyperparameter tuning are performed on the obtained second target prediction model, wherein, during one round of hyperparameter tuning, the following operations are performed: based on the cross-validation value of the second target prediction model, the hyperparameters of the second target prediction model are adjusted, and when the adjusted second target prediction model meets the preset second verification condition, the second learning rate of the second target prediction model is reduced.

[0131] Based on the same inventive concept as the above-mentioned embodiment, the embodiment of the present application further provides an electronic device that can be used for real vehicle attribute prediction. In one embodiment, the electronic device can be a server, or a terminal device or other electronic device. In this embodiment, the structure of the electronic device can be as follows: Figure 7As shown, it includes a memory 701 , a communication interface 703 and one or more processors 702 .

[0132] Memory 701 is used to store computer programs executed by processor 702. Memory 701 may mainly include a program storage area and a data storage area. The program storage area may store an operating system and programs required for running instant messaging functions, while the data storage area may store various instant messaging messages and operating instruction sets.

[0133] Memory 701 may be a volatile memory, such as random-access memory (RAM); a non-volatile memory, such as read-only memory, flash memory, a hard disk drive (HDD), or a solid-state drive (SSD); or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but is not limited thereto. Memory 701 may be a combination of the above memories.

[0134] The processor 702 may include one or more central processing units (CPUs) or digital processing units, etc. The processor 702 is configured to implement the above-mentioned real vehicle attribute prediction method when calling the computer program stored in the memory 701 .

[0135] The communication interface 703 is used to communicate with terminal devices and other servers.

[0136] The specific connection medium between the memory 701, the communication interface 703 and the processor 702 is not limited in the embodiment of the present application. Figure 7 In the embodiment, the memory 701 and the processor 702 are connected via a bus 704. Figure 7 The connections between the other components are shown in bold lines, which are only for illustration and are not intended to be limiting. The bus 704 can be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, Figure 7 Only one thick line is used in the diagram, but this does not mean that there is only one bus or one type of bus.

[0137] Based on the same inventive concept, an embodiment of the present application further provides a storage medium storing computer instructions. When the computer instructions are executed on a computer, the computer executes a real vehicle attribute prediction method discussed above.

[0138] Based on the same inventive concept, an embodiment of the present application further provides a computer program product, which executes a real vehicle attribute prediction method discussed above.

[0139] It should be noted that although several units or subunits of the device are mentioned in the detailed description above, this division is merely exemplary and not mandatory. In fact, depending on the embodiment of the application, the features and functions of two or more units described above can be embodied in a single unit. Conversely, the features and functions of a single unit described above can be further divided and embodied by multiple units.

[0140] Furthermore, although the operations of the method of the present application are described in a particular order in the accompanying drawings, this does not require or imply that the operations must be performed in this particular order, or that all illustrated operations must be performed to achieve the desired results. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step, and / or one step may be decomposed into multiple steps.

[0141] The embodiments of the present application provide a method, device, electronic device and storage medium for predicting the properties of a real vehicle. A preset tire design database and a tire external characteristic database are used to train a corresponding first target prediction model, and a tire external characteristic database and a real vehicle dynamic property database are used to train a corresponding second target prediction model. The tire design database and the tire external characteristic database respectively record the design parameters and dynamic external characteristic parameters of the vehicle tire under real conditions. Based on the above method, the first target prediction model can output a high-precision tire external characteristic prediction result based on the input tire parameters, and enable the second target prediction model tire to further accurately predict the real vehicle properties at the vehicle level based on the tire external characteristic prediction result, thereby effectively improving the prediction accuracy of the real vehicle properties. At the same time, the prediction method based on the above machine learning ensures the prediction efficiency of the real vehicle properties, thereby further improving the production efficiency of vehicle development.

[0142] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0143] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a server, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0144] The program code used to perform the operations of the present application may be written using any combination of one or more programming languages, including object-oriented programming languages ​​such as Java, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computing device, partially on the user's device, as a stand-alone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0145] Where a remote computing device is involved, the remote computing device may be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., through the Internet using an Internet service provider).

[0146] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0147] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0148] Obviously, those skilled in the art may make various changes and modifications to this application without departing from the spirit and scope of this application. Thus, if these modifications and variations of this application fall within the scope of the claims of this application and their equivalents, this application is intended to include these modifications and variations.

Claims

1. A method for predicting real vehicle attributes, characterized in that: include: Using a tire design database and a tire external characteristic database, a preset first prediction model is trained to obtain a corresponding first target prediction model; Using the tire external characteristic database and the real vehicle dynamic property database, a preset second prediction model is trained to obtain a corresponding second target prediction model; Inputting tire parameters of a target tire into the trained first target prediction model to obtain a tire external characteristic prediction result output by the first target prediction model based on the tire parameters; Inputting the tire external characteristic prediction result into the trained second target prediction model to obtain the actual vehicle attribute prediction result output by the second target prediction model; The tire design database includes a plurality of tire parameters, wherein the tire parameters include at least: hysteresis loss DMA of the rubber formula, raw material material, raw material strength, raw material density, raw material damping, parameters related to tire section and inner cavity design, and parameters related to tire pattern design; The tire external characteristic database includes a plurality of tire external characteristic data, the tire external characteristic data including at least rolling resistance, tire weight, contact patch, static stiffness in the X / Y / Z directions, odor, pattern radiation noise, structure-borne vibration noise, transient impact vibration, dynamic stiffness in the X / Y / Z directions, moment in the X / Y / Z directions, relaxation length, force transfer rate, and pattern drainage; or the tire external characteristic data includes at least rolling grip, tire weight, cornering stiffness, tire groove depth, rubber viscoelasticity, geometric cross-sectional dimensions, vertical dynamic stiffness, longitudinal slip coefficient, and dynamic load radius; The real vehicle dynamics attribute database contains a variety of real vehicle dynamics attribute data, and the real vehicle dynamics attribute data at least includes: vehicle dynamic handling stability, vehicle comfort, vehicle quietness, dynamic model chassis algorithm model intervention, chassis integrated algorithm model intervention, vehicle medium and low adhesion road surface and electronic control system safety.

2. The method according to claim 1, wherein Before training the preset first prediction model using the tire design database and the tire external characteristic database, the method further includes: performing data cleaning on the obtained tire design data using a cleaning rule, and constructing a corresponding tire design database, wherein the tire design data includes a design parameter of at least one vehicle tire and a parameter range determined for the design parameter of the at least one vehicle tire; Using the cleaning rule, the obtained tire external characteristic data is cleaned and a corresponding tire external characteristic database is constructed, wherein the tire external characteristic data includes at least one external characteristic parameter of a vehicle tire; The cleaning rules are used to clean the acquired real vehicle dynamics attribute data and construct a corresponding real vehicle dynamics attribute database, wherein the real vehicle dynamics attribute database includes the real vehicle dynamics attributes of at least one vehicle and attribute ranges determined for the real vehicle dynamics attributes of the at least one vehicle.

3. The method according to claim 2, wherein The tire design database and the tire external characteristic database are used to train a preset first prediction model to obtain a corresponding first target prediction model, including: Using a tire design database and a tire external characteristic database, setting a corresponding first training sample set, wherein one first training sample includes: first input information determined based on the tire design database and first standard information determined based on the tire external characteristic database; Using the first training sample in the first training sample set, a preset first prediction model is trained for multiple rounds of iterative training, and when a preset first convergence condition is met, a first target prediction model is output; wherein, during one round of iterative training, the following operations are performed: The first prediction model is used to obtain a corresponding external characteristic prediction result based on the first input information in the first training sample, and the parameters of the first prediction model are adjusted based on the loss value between the external characteristic prediction result and its corresponding first standard information.

4. The method according to claim 3, wherein After obtaining the corresponding first target prediction model, the method further includes: Using a preset first learning rate, multiple rounds of hyperparameter tuning are performed on the obtained first target prediction model, wherein in one round of hyperparameter tuning, the following operations are performed: Based on the cross-validation value of the first target prediction model, the hyperparameters of the first target prediction model are adjusted, and when the adjusted first target prediction model meets a preset first verification condition, the first learning rate of the first target prediction model is reduced.

5. The method according to any one of claims 1 to 3, wherein The method of using the tire external characteristic database and the real vehicle dynamics attribute database to train a preset second prediction model to obtain a corresponding second target prediction model includes: Using the tire external characteristic database and the actual vehicle dynamic property database, a corresponding second training sample set is set, wherein one second training sample includes: second input information determined based on the tire external characteristic database and second standard information determined based on the actual vehicle dynamic property database; Using the second training samples in the second training sample set, performing multiple rounds of iterative training on the preset second prediction model, and outputting a second target prediction model when a preset second convergence condition is met; wherein, during one round of iterative training, the following operations are performed: The second prediction model is used to obtain a corresponding real vehicle attribute prediction result based on the second input information in the second training sample, and the parameters of the second prediction model are adjusted based on the loss value between the real vehicle attribute prediction result and its corresponding second standard information.

6. The method according to claim 5, wherein After obtaining the corresponding second target prediction model, the method further includes: Using a preset second learning rate, multiple rounds of hyperparameter tuning are performed on the obtained second target prediction model, wherein in one round of hyperparameter tuning, the following operations are performed: Based on the cross-validation value of the second target prediction model, the hyperparameters of the second target prediction model are adjusted, and when the adjusted second target prediction model meets a preset second verification condition, the second learning rate of the second target prediction model is reduced.

7. A real vehicle attribute prediction device, characterized in that: include: A first training module is used to train a preset first prediction model using a tire design database and a tire external characteristic database to obtain a corresponding first target prediction model; A second training module is configured to train a preset second prediction model using the tire external characteristic database and the real vehicle dynamics attribute database to obtain a corresponding second target prediction model; an external characteristic prediction module, configured to input tire parameters of a target tire into the trained first target prediction model, and obtain a tire external characteristic prediction result output by the first target prediction model based on the tire parameters; an attribute prediction module, configured to input the tire external characteristic prediction result into the trained second target prediction model to obtain a real vehicle attribute prediction result output by the second target prediction model; The tire design database includes a plurality of tire parameters, wherein the tire parameters include at least: hysteresis loss DMA of the rubber formula, raw material material, raw material strength, raw material density, raw material damping, parameters related to tire section and inner cavity design, and parameters related to tire pattern design; The tire external characteristic database includes a plurality of tire external characteristic data, the tire external characteristic data including at least rolling resistance, tire weight, contact patch, static stiffness in the X / Y / Z directions, odor, pattern radiation noise, structure-borne vibration noise, transient impact vibration, dynamic stiffness in the X / Y / Z directions, moment in the X / Y / Z directions, relaxation length, force transfer rate, and pattern drainage; or the tire external characteristic data includes at least rolling grip, tire weight, cornering stiffness, tire groove depth, rubber viscoelasticity, geometric cross-sectional dimensions, vertical dynamic stiffness, longitudinal slip coefficient, and dynamic load radius; The real vehicle dynamics attribute database contains a variety of real vehicle dynamics attribute data, and the real vehicle dynamics attribute data at least includes: vehicle dynamic handling stability, vehicle comfort, vehicle quietness, dynamic model chassis algorithm model intervention, chassis integrated algorithm model intervention, vehicle medium and low adhesion road surface and electronic control system safety.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable by the processor, wherein: When the processor executes the computer program, the method according to any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product, characterized in that When the computer program product is executed, the method according to any one of claims 1 to 6 is implemented.

Images