Modeling method, device and equipment for tire dynamic model and medium

Abstract

The method is mainly applied to the technical field of vehicle engineering. The invention discloses a tire dynamics model modeling method, device and equipment and a medium, and the method comprises the steps: carrying out the parameter initialization of a tire dynamics basic model based on the test data of the dynamics performance of a tire, and determining a basic prediction value through the initialized basic model; using a preset neural network model to learn an error distribution rule of the test data and the basic prediction value under a complex working condition, and outputting an error compensation value corresponding to the basic prediction value; fusing the basic predicted value and the error compensation value by adopting a dynamic weighted fusion strategy to obtain a preliminary fusion tire force; and carrying out physical rationality correction on the preliminarily fused tire force to obtain the tire force meeting physical constraint conditions. According to the method, the tire model with high precision, high adaptability and high robustness can be constructed under complex working conditions.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of vehicle engineering, in particular to a tire dynamics model modeling method, device, equipment and medium. BACKGROUND

[0002] With the continuous evolution of vehicle dynamics simulation accuracy towards extreme working conditions and multi-physical field coupling, the modeling quality of the tire, as the only load-bearing and force transmission component directly contacting the ground, has become a key bottleneck determining the credibility of the whole vehicle simulation. The current mainstream tire dynamics model can be roughly divided into two categories: the empirical model represented by Magic Formula (MF) is a pure mathematical fitting of steady-state test data through algebraic functions, with high working condition adaptability and calculation efficiency; the physical model establishes mechanical differential equations based on contact mechanics and sliding friction theory, striving to explain the physical nature of force / torque generation under the premise of structural unity.

[0003] However, a single model architecture often finds it difficult to balance physical consistency, working condition generality and structural unity, forming an inherent triangular contradiction. First, physical consistency requires the model to correctly reflect the nonlinear coupling mechanisms of friction limit, dynamic slip, temperature, load distribution, etc. If the empirical model pursues higher accuracy, it must add multiple layers of local correction terms to the original function architecture, resulting in a sharp increase in the number of parameters and mutual coupling, ultimately making the model lose clear physical explanation; while the pure physical model often based on idealized Brush or Fiala assumptions, needs to introduce additional empirical corrections in the medium and large slip regions and complex slip working conditions, otherwise the prediction error is significant, and the physical consistency is destroyed at the boundary.

[0004] Second, working condition generality requires the same model to cover all working conditions from low-speed parking, high-speed cruising to extreme drifting, off-road obstacles, etc. The empirical model relies on large-scale calibration data, and the generalization ability for different road surfaces, tire structures or environmental conditions is limited by data distribution; if the same set of parameters or function form is used, the prediction deviation is often amplified sharply in the training data sparse area. The physical model theoretically has stronger extrapolation ability, but the overly simplified unified structure cannot capture the details of the multi-layer materials inside the tire, non-uniform pressure distribution, etc., and the generality is restricted by the structural unity.

[0005] Finally, in order to balance physical consistency and universality, the model has to introduce segmented functions, interpolation tables or mixed logic, so that the originally unified mathematical form evolves into a "quasi-unified" or "multi-mode" architecture, losing the theoretical simplicity and numerical robustness. In summary, the traditional single tire model inevitably appears the parameter conflict between physical consistency and structural unity, the functional expansion pressure of working condition universality on structural unity, and the data dependency of physical consistency on working condition universality under the triple constraints, which mutually restrict each other and constitute a systematic dilemma that the existing technology cannot break through. Therefore, a new modeling method is needed, which can realize the coordination of the three at a higher level. SUMMARY

[0006] The application provides a tire dynamics model modeling method, device, equipment and medium, which can construct a high-precision, high-adaptability and high-robustness tire model under complex working conditions.

[0007] The application provides a tire dynamics model modeling method, which comprises the following steps: Based on the test data of the dynamic performance of the tire, the parameters of the basic model of tire dynamics are initialized, and the basic prediction value is determined by using the initialized basic model; The preset neural network model is used to learn the error distribution law of the test data and the basic prediction value under complex working conditions, and output the error compensation value corresponding to the basic prediction value; The dynamic weighted fusion strategy is adopted to fuse the basic prediction value and the error compensation value, and the preliminary fused tire force is obtained; The physical rationality of the preliminary fused tire force is corrected to obtain the tire force that meets the physical constraint condition.

[0008] Optionally, based on the test data of the dynamic performance of the tire, the parameters of the basic model of tire dynamics are initialized, and the basic prediction value is determined by using the initialized basic model, which comprises the following steps: The magic formula is selected as the core framework of the basic model, and the magic formula adopts a trigonometric function structure to describe the basic variation trend of the tire force; The longitudinal slip ratio, the side slip angle and the vertical load in the test data are taken as the input variables of the magic formula, and after operation through the magic formula, the initial longitudinal force and the initial lateral force are output as the first basic prediction value and the second basic prediction value respectively.

[0009] Optionally, the preset neural network model is used to learn the error distribution law of the test data and the basic prediction value under complex working conditions, and output the error compensation value corresponding to the basic prediction value, which comprises the following steps: taking a first error of the initial longitudinal force relative to a measured tire force and a second error of the initial lateral force relative to the measured tire force as a training target of the neural network model; taking a coupling term of a basic input variable, an intermediate output variable of the magic formula, and a working condition state identifier representing a current vehicle dynamics state as an input feature of the neural network model; training the neural network model based on the training target and the input feature, so that the neural network model learns the error distribution rule and outputs a longitudinal force compensation and a lateral force compensation corresponding to the basic prediction value under complex working conditions.

[0010] Optionally, the dynamic weighting fusion strategy is adopted to fuse the basic prediction value and the error compensation value to obtain a preliminary fused tire force, including: dynamically generating an adaptive fusion weight proportion value of the basic prediction value and the error compensation value based on a current vehicle motion state and a road adhesion condition; weighting and fusing the basic prediction value and the error compensation value according to the adaptive fusion weight proportion value to obtain a fused tire force prediction value as the preliminary fused tire force.

[0011] Optionally, the adaptive fusion weight proportion value of the basic prediction value and the error compensation value is dynamically generated based on a current vehicle motion state and a road adhesion condition, including: determining a current working condition according to a longitudinal slip ratio and a lateral side slip angle; if the current working condition is a regular working condition, assigning a first weight value in a first numerical interval to the basic prediction value and a second weight value in the first numerical interval to the error compensation value, wherein the first weight value is greater than the second weight value; if the current working condition is a complex working condition, assigning a third weight value in a second numerical interval to the basic prediction value and a fourth weight value in the second numerical interval to the error compensation value, wherein the third weight value is less than the fourth weight value; the preliminary fused tire force is equal to the basic prediction value multiplied by the first weight value or the third weight value, plus the error compensation value multiplied by the second weight value or the fourth weight value.

[0012] Optionally, the preliminary fused tire force is physically reasonable, including: inputting the fused tire force prediction value into a constraint model based on tire physical characteristics for physical reasonableness verification; When it is determined according to the result output by the constraint model that the fused tire force prediction value violates a physical constraint condition, the tire force prediction value satisfying the physical constraint condition is output after the fused tire force prediction value is adjusted.

[0013] Optionally, the constraint model is a constraint model based on a friction circle theory, and the basic prediction value includes a longitudinal force prediction value and a lateral force prediction value. The manner of adjusting the fused tire force prediction value includes: calculating a resultant force value of the preliminary fused tire force; comparing the resultant force value with a maximum allowable friction force calculated based on a current road adhesion coefficient and a vertical load; when the resultant force value is greater than the maximum allowable friction force, reducing the longitudinal force prediction value and the lateral force prediction value by a preset proportion until the resultant force value is equal to or less than the maximum allowable friction force.

[0014] The application further provides a tire dynamics model modeling device, the device comprising: a first operation module configured to initialize a basic model of tire dynamics based on test data of dynamic performance of a tire, and determine a basic prediction value by using the initialized basic model; a second operation module configured to learn error distribution rules of the test data and the basic prediction value under complex working conditions by using a preset neural network model, and output an error compensation value corresponding to the basic prediction value; a fusion module configured to fuse the basic prediction value and the error compensation value by using a dynamic weighted fusion strategy to obtain a preliminary fused tire force; an adjustment module configured to correct the preliminary fused tire force for physical reasonableness to obtain a tire force satisfying a physical constraint condition.

[0015] The application further provides an electronic device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the tire dynamics model modeling method according to any one of the above when executing the computer program.

[0016] The application further provides a computer readable storage medium storing a computer program, wherein the computer program is executed by a processor to implement the tire dynamics model modeling method according to any one of the above.

[0017] The application has at least the following beneficial effects: Firstly, the parameter initialization of the basic model is carried out based on the tire dynamic performance test data, and the basic prediction value is determined, which provides an initial basis for model construction. Secondly, the error distribution law of the test data and the basic prediction value under complex working conditions is learned by using a neural network, and an error compensation value is output. This step can effectively capture the subtle change law of the tire force under complex working conditions by means of the powerful nonlinear fitting capability of the neural network, and improve the adaptability of the model to complex working conditions. Then, a dynamic weighted fusion strategy is adopted to fuse the basic prediction value and the error compensation value, and the dynamic adjustment of the weight can make the fusion result more consistent with the actual working condition, and further improve the model precision. Finally, the tire force after preliminary fusion is physically corrected to ensure that the tire force meets the physical constraint condition, and the robustness of the model is enhanced, so that reliable results can be stably output under different complex working conditions. BRIEF DESCRIPTION OF DRAWINGS

[0018] The accompanying drawings are used to provide a further understanding of the technical solutions of the present application, and constitute a part of the specification, and are used to explain the technical solutions of the present application together with the embodiments of the present application, and do not constitute a limitation on the technical solutions of the present application.

[0019] Figure 1 It is a step flow chart of a tire dynamics model modeling method; Figure 2 It is a step flow chart of step S101 in the tire dynamics model modeling method; Figure 3 It is a step flow chart of step S102 in the tire dynamics model modeling method; Figure 4 It is a step flow chart of step S103 in the tire dynamics model modeling method; Figure 5 It is a step flow chart of step S104 in the tire dynamics model modeling method; Figure 6 It is a structural schematic diagram of a tire dynamics model modeling device; Figure 7 It is a step flow chart of the running process of the tire dynamics model modeling device; Figure 8 It is a structural schematic diagram of an electronic device. DETAILED DESCRIPTION

[0020] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, and do not limit the present application.

[0021] It should be noted that in the field of vehicle dynamics simulation and control, the tire model is a key link connecting the vehicle and the road, and its modeling method mainly includes three core technical paths: (1) Semi-empirical tire model: represented by the Magic Formula (MF) tire model, this model fits tire test data through an empirical formula composed of trigonometric functions, and can describe the nonlinear relationship between key mechanical properties such as longitudinal force and lateral force and input parameters such as slip ratio and side slip angle. Due to its strong engineering practicality, it is widely used in industrial simulation. Similar technologies include the Dugoff model and the Fiala brush model based on the friction circle assumption. These models simplify the tire contact mechanics mechanism and construct mathematical expressions with clear physical meaning, but the model parameters are difficult to obtain, the adaptability to complex working conditions is insufficient, and the model expansion ability is limited.

[0022] (2) Pure data-driven model: with the development of deep learning technology, neural network-based tire models have become a research hotspot. This type of model uses the universal approximation characteristics of neural networks to directly learn the mapping relationship between input and output from test data, without relying on complex physical mechanism derivation, and naturally has an advantage in nonlinear fitting, but pure data-driven models lack physical constraints, have high data dependence, and have poor interpretability.

[0023] (3) Hybrid modeling technology: some studies attempt to combine the physical constraints of mechanism models with the learning ability of data-driven methods, such as the neural ordinary differential equation (NODE) tire model, which optimizes the modeling effect of neural networks by incorporating friction ellipse constraints and "S-shaped" saturation trends, but hybrid models mostly only embed simple physical constraints in neural networks, and do not fully utilize the structural advantages of mechanism models, and do not achieve deep integration of mechanism knowledge and data information, and have not yet achieved an optimal balance between full-working-condition adaptability and modeling efficiency.

[0024] Please refer to Figure 1 , Figure 1 is a step flowchart of a tire dynamics model modeling method.

[0025] The embodiment provides a tire dynamics model modeling method, which comprises: S101, based on the test data of the dynamic performance of the tire, the parameters of the basic model of the tire dynamics are initialized, and the basic prediction value is determined by using the initialized basic model.

[0026] S102, using a preset neural network model, learning the error distribution law of the test data and the basic prediction value under complex working conditions, and outputting the error compensation value corresponding to the basic prediction value.

[0027] S103, the preliminary tire force is obtained by fusing the basic prediction value and the error compensation value by using a dynamic weighting fusion strategy.

[0028] S104, the tire force satisfying the physical constraint condition is obtained by performing physical rationality correction on the preliminary tire force.

[0029] It can be understood that in the embodiment, first, the basic model is parameterized and initialized based on the tire dynamic performance test data to determine the basic prediction value, providing an initial basis for model construction. Second, the error distribution law of the test data and the basic prediction value under complex working conditions is learned by using a neural network to output an error compensation value. This step can effectively capture the subtle change law of the tire force under complex working conditions by virtue of the powerful nonlinear fitting capability of the neural network, thereby improving the adaptability of the model to complex working conditions. Third, the basic prediction value and the error compensation value are fused by using a dynamic weighting fusion strategy, and dynamic adjustment of the weight can make the fusion result more consistent with the actual working condition, thereby further improving the model precision. Finally, the tire force is physically corrected to ensure that the tire force satisfies the physical constraint condition, thereby enhancing the robustness of the model and enabling it to stably output reliable results under different complex working conditions.

[0030] Please refer to Figure 2 , Figure 2 is a step flowchart of step S101 in a tire dynamics model modeling method.

[0031] In some embodiments, step S101 comprises: S201, a magic formula is selected as the core framework of the basic model, and the magic formula uses a trigonometric function structure to describe the basic change trend of the tire force.

[0032] S202, the longitudinal slip ratio, the side slip angle and the vertical load in the test data are taken as the input variables of the magic formula, and after operation by the magic formula, the initial longitudinal force and the initial lateral force are output as the first basic prediction value and the second basic prediction value respectively.

[0033] It can be understood that the magic formula is a mathematical model widely used in tire modeling, and its core is to use a trigonometric function structure to describe the basic change trend of the tire force. This formula can well simulate the longitudinal force and the lateral force of the tire under different working conditions. Taking the magic formula as the basic framework of the tire model ensures that the model can capture the nonlinear relationship between the tire force and the key parameters such as the slip ratio and the side slip angle.

[0034] In some embodiments, test data of the tire under different working conditions, including longitudinal slip ratio, side slip angle and vertical load, are collected in laboratory or actual road test. For example, the longitudinal slip ratio can vary between 0% and 20%, the side slip angle can vary between 0° and 10°, and the vertical load can vary between 300 kg and 800 kg.

[0035] The longitudinal slip ratio, the side slip angle and the vertical load in the test data are taken as input variables of the magic formula. The magic formula operates on these input variables through an internal trigonometric function structure.

[0036] The magic formula outputs an initial longitudinal force and an initial lateral force after operation, which are taken as a first basic prediction value and a second basic prediction value, respectively. For example, when the longitudinal slip ratio is 10%, the side slip angle is 5°, and the vertical load is 500 kg, the magic formula can output an initial longitudinal force of 1000 N and an initial lateral force of 500 N.

[0037] It can be understood that, by selecting the magic formula as the core framework of the basic model, the basic change trend of the tire force is accurately described by using the trigonometric function structure, and a solid theoretical basis is provided for the model. At the same time, the longitudinal slip ratio, the side slip angle and the vertical load are taken as input variables, so that the model can more comprehensively reflect the force characteristics of the tire under different working conditions. Combined with subsequent neural network learning, dynamic weighted fusion and physical rationality correction, the tire model constructed under complex working conditions has higher precision, stronger adaptability and better robustness, and can more accurately predict the tire force to meet the actual application requirements.

[0038] Please refer to Figure 3 , Figure 3 is a step flowchart of step S102 in a tire dynamics model modeling method.

[0039] In some embodiments, step S102 comprises: S301, taking a first error of the initial longitudinal force relative to the measured tire force and a second error of the initial lateral force relative to the measured tire force as a training target of the neural network model.

[0040] S302, taking a coupling term of the basic input variable, an intermediate output variable of the magic formula and a working condition state identifier representing the current vehicle dynamics state as an input feature of the neural network model.

[0041] S303, training the neural network model based on the training target and the input feature, so that the neural network model learns the error distribution law and outputs a longitudinal force compensation amount and a lateral force compensation amount corresponding to the basic prediction value under complex working conditions.

[0042] In some embodiments, the coupling term of the basic input variables (longitudinal slip ratio, side slip angle, vertical load) is taken as one of the input features. The coupling term can reflect the mutual influence between these variables.

[0043] In some embodiments, the intermediate output variables generated by the magic formula during the operation are taken as one of the input features. These intermediate variables may contain potential information of tire force changes.

[0044] In some embodiments, the working condition state identifier (such as road surface type, vehicle speed, etc.) representing the current vehicle dynamics state is taken as one of the input features to enhance the adaptability of the model to complex working conditions.

[0045] It can be understood that, by taking the error of the initial longitudinal force and lateral force as the training target, the neural network can accurately learn the error distribution rule of the tire force. At the same time, by introducing the basic input variable coupling term, the magic formula intermediate output variable, and the working condition state identifier as input features, the neural network can more comprehensively capture the tire force changes under complex working conditions. The trained neural network outputs the compensation amount of the longitudinal force and the lateral force, further corrects the basic prediction value, and significantly improves the accuracy and robustness of the model under complex working conditions, making it more suitable for actual vehicle dynamics requirements.

[0046] Please refer to Figure 4 , Figure 4 is a step flowchart of step S103 in a tire dynamics model modeling method.

[0047] In some embodiments, step S103 comprises: S401, based on the current vehicle motion state and the road adhesion condition, dynamically generating an adaptive fusion weight proportion value of the basic prediction value and the error compensation value.

[0048] S402, weighting and fusing the basic prediction value and the error compensation value according to the adaptive fusion weight proportion value to obtain a fused tire force prediction value and take it as the preliminary fused tire force.

[0049] In some embodiments, step S401 comprises: The current working condition is determined according to the longitudinal slip ratio and the lateral side slip angle; if the current working condition is a normal working condition, a first weight value in a first numerical interval is assigned to the basic prediction value, and a second weight value in the first numerical interval is assigned to the error compensation value, wherein the first weight value is greater than the second weight value; if the current working condition is a complex working condition, a third weight value in a second numerical interval is assigned to the basic prediction value, and a fourth weight value in the second numerical interval is assigned to the error compensation value, wherein the third weight value is less than the fourth weight value; the preliminary fused tire force is equal to the basic prediction value multiplied by the first weight value or the third weight value, plus the error compensation value multiplied by the second weight value or the fourth weight value.

[0050] It can be understood that, by dynamically generating adaptive fusion weight proportion values according to the vehicle motion state and the road adhesion condition, the model can adjust the weight distribution of the basic prediction value and the error compensation value in real time. This dynamic adjustment mechanism makes the model more flexible under different working conditions, and can more accurately reflect the actual stress of the tire. Combined with the previous magic formula initialization, neural network error compensation and physical rationality correction, the tire model constructed under complex working conditions has higher precision, stronger adaptability and better robustness, and can better meet the diversified needs in actual applications.

[0051] Please refer to Figure 5 , Figure 5 A step flowchart of step S104 in a tire dynamics model modeling method.

[0052] In some embodiments, step S104 comprises: S501, input the fused tire force prediction value to a constraint model constructed based on tire physical characteristics for physical rationality verification.

[0053] S502, when it is determined according to the output result of the constraint model that the fused tire force prediction value violates the physical constraint condition, the fused tire force prediction value is adjusted, and the tire force prediction value satisfying the physical constraint condition is output.

[0054] In some embodiments, the constraint model is a constraint model based on the friction circle theory, and the basic prediction value includes a longitudinal force prediction value and a lateral force prediction value.

[0055] In some embodiments, the way of adjusting the fused tire force prediction value comprises: calculating the resultant force value of the preliminary fused tire force; comparing the resultant force value with the maximum allowable friction force calculated based on the current road adhesion coefficient and the vertical load; when the resultant force value is greater than the maximum allowable friction force, reducing the longitudinal force prediction value and the lateral force prediction value by a preset proportion until the resultant force value is equal to or less than the maximum allowable friction force.

[0056] It can be understood that the embodiment inputs the fused tire force prediction value into the constraint model based on the tire physical characteristics to verify, ensures that the tire force prediction value output by the model conforms to the physical law. When the prediction value violates the physical constraint condition, timely adjustment and correction are made to avoid the occurrence of unreasonable results. Combined with the previous dynamic weighted fusion, magic formula initialization, neural network error compensation and other means, the tire model constructed under complex working conditions has higher precision, stronger adaptability, and can maintain stable output under extreme or abnormal working conditions, significantly improving the practicality and reliability of the model.

[0057] Please refer to Figure 6 , Figure 6 It is a structural schematic diagram of a tire dynamics model modeling device.

[0058] The embodiment also provides a tire dynamics model modeling device comprising: A first operation module 601 is configured to initialize parameters of a basic model of tire dynamics based on test data of dynamic performance of the tire, and determine a basic prediction value by using the initialized basic model; A second operation module 602 is configured to learn error distribution rules of the test data and the basic prediction value under complex working conditions by using a preset neural network model, and output an error compensation value corresponding to the basic prediction value; A fusion module 603 is configured to fuse the basic prediction value and the error compensation value by using a dynamic weighted fusion strategy to obtain a preliminary fused tire force; An adjustment module 604 is configured to perform physical rationality correction on the preliminary fused tire force to obtain a tire force meeting a physical constraint condition.

[0059] Please refer to Figure 7 , Figure 7 It is a step flowchart of a running process of a tire dynamics model modeling device.

[0060] The technical solution provides a specific embodiment, and the tire dynamics model modeling device of the embodiment is composed of four parts of a magic formula basic module, a neural network compensation module, a multi-scale data fusion module and a physical constraint verification module. The running process is as follows: First, a classic magic formula (such as Pacejka 2002) is selected as a core framework, the trigonometric function structure for describing the basic change trend of the tire force is retained, and input parameters include key variables such as longitudinal slip ratio κ, side slip angle α, vertical load Fz and the like, and output initial longitudinal force Fx0 and lateral force Fy0. By simplifying the parameter calibration process, the basic parameter initialization is completed by using conventional test data (such as pure longitudinal slip and pure side slip working condition data), and the basic accuracy and physical rationality of the model under conventional working conditions are ensured.

[0061] Secondly, a light-weight feedforward neural network (input layer - hidden layer - output layer) is designed, with the prediction error of the magic formula (ΔFx = Fx_measured - Fx0, ΔFy = Fy_measured - Fy0) as the training target, and the input features including the coupling terms of the basic input variables (κ α, κ², α², etc.), capturing the coupling effects of combined slip conditions.

[0062] The intermediate output of the magic formula (such as the shape factor, peak force, etc.), strengthening the relevance of the mechanism and data.

[0063] The condition state identifier (such as whether it is close to the adhesion limit, whether it is a non-steady-state excitation), improving the compensation accuracy of complex conditions.

[0064] The neural network learns the error distribution law of the magic formula under complex conditions, and outputs accurate compensation amounts ΔFx*, ΔFy*.

[0065] Then, a weighted fusion strategy is used to integrate the outputs of the basic module and the compensation module, and the fusion weight is adjusted adaptively by the condition complexity: Under normal conditions (small slip rate, small side slip angle), increase the weight of the magic formula (0.7-0.9) to ensure the calculation efficiency and physical stability of the model; under complex conditions (large slip rate, combined slip, close to the adhesion limit), increase the weight of the neural network (0.6-0.8) to fully exert the nonlinear fitting advantages of data-driven.

[0066] The final tire force output is: Fx = Fx0 + w ΔFx*, Fy = Fy0 + w ΔFy* (w is the adaptive weight).

[0067] Finally, a physical constraint verification link is added at the output end of the model, based on the friction circle theory:

[0068] Where μ is the road adhesion coefficient) to modify the fusion result, and if the output exceeds the physical limit, it is adjusted to a reasonable range through a saturation function, ensuring that the model always meets the basic constraints of tire dynamics and avoids non-physical output.

[0069] Those of ordinary skill in the art will understand that all or some of the steps, devices, and apparatuses in the methods disclosed above can be implemented as software, firmware, hardware, or a suitable combination thereof. Some or all of the physical components can be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or as hardware, or as an integrated circuit, such as an application specific integrated circuit. Such software can be distributed on computer readable media, which can include computer storage media (or non-transitory media) and communication media (or transitory media). As is well known to those of ordinary skill in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and that can be accessed by a computer. As is well known to those of ordinary skill in the art, communication media typically embodies computer readable instructions, data structures, program modules or other data in a modulated data signal such as a carrier wave or other transport mechanism and includes any information delivery media.

[0070] It can be understood that the contents of the above method embodiments are applicable to the device embodiments, the device embodiments specifically implement the functions of the above method embodiments, and achieve the same beneficial effects as the above method embodiments.

[0071] The embodiments of the present application further provide an electronic device, which comprises a memory and a processor. The memory stores a computer program. When the processor executes the computer program, the tire dynamics modeling method according to any one of the above embodiments is implemented.

[0072] Reference Figure 8 , Figure 8 The hardware structure of the electronic device of another embodiment is illustrated, which comprises: The processor 701 can be implemented in a general-purpose CPU (Central Processing Unit), a microprocessor, an ASIC (Application Specific Integrated Circuit), or one or more integrated circuits, and is used to execute related programs to implement the technical solutions provided by the embodiments of the present application. The memory 702 can be implemented in the form of Read Only Memory (ROM), static storage device, dynamic storage device or Random Access Memory (RAM), etc. The memory 702 can store operating devices and other application programs, and when the technical solutions provided by the embodiments of the present specification are implemented by software or firmware, the related program codes are stored in the memory 702 and are called and executed by the processor 701 to implement the tire dynamics model modeling method of the embodiments of the present application; The input / output interface 703 is used to realize information input and output; The communication interface 704 is used to realize the communication interaction between the device and other devices, which can realize communication through wired mode (such as USB, network cable, etc.) or wireless mode (such as mobile network, WIFI, Bluetooth, etc.); The bus 705 transmits information between various components (such as the processor 701, the memory 702, the input / output interface 703 and the communication interface 704) of the device; The processor 701, the memory 702, the input / output interface 703 and the communication interface 704 realize the communication connection between each other in the device through the bus 705.

[0073] It can be understood that the contents in the above method embodiments are all applicable to the electronic device embodiments, the functions specifically implemented by the electronic device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those of the above method embodiments.

[0074] The embodiments of the present application also provide a computer readable storage medium, which stores a program executable by a processor, and the processor executes the program to implement the tire dynamics model modeling method according to any one of the above embodiments.

[0075] The embodiments of the present application also disclose a computer program product, which includes a computer program or computer instructions, the computer program or computer instructions are stored in a computer readable storage medium, a processor of a computer device reads the computer program or computer instructions from the computer readable storage medium, and the processor executes the computer program or computer instructions to make the computer device execute the tire dynamics model modeling method according to any one of the above embodiments.

[0076] It can be understood that the contents in the above method embodiments are all applicable to the storage medium embodiments, the functions specifically implemented by the storage medium embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those of the above method embodiments.

[0077] The terms "first", "second", "third", "fourth", and the like in the description of the specification and in the claims of the present application, if any, are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the use of the terms so construed herein is merely for convenience and brevity and that one of ordinary skill in the art will be able to devise examples of the present application with the features of the application in different orders and / or configurations without departing from the spirit or ambit of the application. Furthermore, the terms "comprise", "comprising", "include", "including", and "has", "having" and variants thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises, has, includes an item or list of items that does not include other non-specified items or remove other non-specified items likewise falls within the scope of the present application. It will be appreciated that the terms "at least one", "one or more", "multiple", "two or more" and the like used herein are meant to encompass a quantity of one or more.

[0078] In several embodiments provided in the present application, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the above-described device embodiments are merely illustrative, and the division of the units is merely a logical function division. In actual implementation, another division manner can be used, for example, a plurality of units or components can be combined or integrated into another device, or some features can be omitted or not executed. In addition, the coupling or direct coupling or communication connection between the units or components shown or discussed can be indirect coupling or communication connection through some interfaces, devices or units, and can be electrical, mechanical or other forms.

[0079] The units described as separate components can or can not be physically separate, and the components shown as units can or can not be physical units, that is, can be located in one place, or can be distributed on a plurality of network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the embodiment.

[0080] In addition, each functional unit in each embodiment of the present application can be integrated into one processing unit, or each unit can be physically present separately, or two or more units can be integrated into one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit.

[0081] The integrated unit, if implemented in the form of a software function unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or the whole or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in the embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various media that can store program codes.

[0082] Although the description of the present application has been quite detailed and particularly described with respect to several described embodiments, it is not intended to be limited to any of these details or embodiments or any special embodiment, but should be considered to effectively cover the intended scope of the present application by referring to the appended claims, taking into account the broadest possible interpretation of these claims in view of the prior art. In addition, the present application is described above in embodiments that the inventors can foresee, and the purpose is to provide a useful description, and non-essential modifications to the present application that have not yet been foreseen can still represent equivalent modifications to the present application.

Claims

1. A tire dynamics model modeling method, characterized by, The method comprises: Parameter initialization is performed on a basic model of tire dynamics based on test data of tire dynamics, and a basic prediction value is determined by using the initialized basic model; A preset neural network model is used to learn error distribution rules of the test data and the basic prediction value under complex working conditions, and an error compensation value corresponding to the basic prediction value is output; A dynamic weighting fusion strategy is used to fuse the basic prediction value and the error compensation value to obtain a preliminary fused tire force; Physical rationality correction is performed on the preliminary fused tire force to obtain a tire force meeting physical constraint conditions.

2. The method of claim 1, wherein, The parameter initialization is performed on the basic model of tire dynamics based on the test data of tire dynamics, and the basic prediction value is determined by using the initialized basic model, and comprises: A magic formula is selected as a core framework of the basic model, and the magic formula uses a trigonometric function structure to describe a basic variation trend of the tire force; Longitudinal slip ratios, side slip angles and vertical loads in the test data are taken as input variables of the magic formula, and initial longitudinal forces and initial lateral forces output after operation of the magic formula are taken as first and second basic prediction values respectively.

3. The method of claim 2, wherein, The preset neural network model is used to learn the error distribution rules of the test data and the basic prediction value under complex working conditions, and the error compensation value corresponding to the basic prediction value is output, and comprises: First and second errors of the initial longitudinal force and the initial lateral force relative to a measured tire force are taken as training targets of the neural network model; Coupling terms of basic input variables, intermediate output variables of the magic formula and working condition state identifiers representing current vehicle dynamics states are taken as input features of the neural network model; The neural network model is trained based on the training targets and the input features, so that the neural network model learns the error distribution rules and outputs longitudinal force compensation amounts and lateral force compensation amounts corresponding to the basic prediction value under complex working conditions.

4. The method according to any one of claims 1 to 3, characterized in that, The dynamic weighting fusion strategy is used to fuse the basic prediction value and the error compensation value to obtain the preliminary fused tire force, and comprises: An adaptive fusion weight proportion value of the basic prediction value and the error compensation value is dynamically generated based on current vehicle motion states and road adhesion conditions; The basic prediction value and the error compensation value are weighted fused according to the adaptive fusion weight proportion value to obtain a fused tire force prediction value as the preliminary fused tire force.

5. The method of claim 4, wherein, The adaptive fusion weight proportion value of the basic prediction value and the error compensation value is dynamically generated based on the current vehicle motion states and the road adhesion conditions, and comprises: A current working condition is determined according to longitudinal slip ratios and lateral side slip angles; If the current working condition is a regular working condition, a first weight value in a first numerical interval is assigned to the basic prediction value, and a second weight value in the first numerical interval is assigned to the error compensation value, wherein the first weight value is greater than the second weight value; if the current working condition is a complex working condition, assigning a third weight value in a second numerical interval to the basic prediction value and a fourth weight value in the second numerical interval to the error compensation value, wherein the third weight value is less than the fourth weight value; the preliminary fused tire force is equal to the basic prediction value multiplied by the first weight value or the third weight value, plus the error compensation value multiplied by the second weight value or the fourth weight value.

6. The method of claim 4, wherein, the physical rationality correction on the preliminary fused tire force to obtain a tire force meeting physical constraints comprises: inputting the fused tire force prediction value into a constraint model based on tire physical characteristics for physical rationality checking; when the result output according to the constraint model determines that the fused tire force prediction value violates the physical constraint condition, adjusting the fused tire force prediction value and outputting a tire force prediction value meeting the physical constraint condition.

7. The method of claim 6, wherein, the constraint model is a constraint model based on the friction circle theory, and the basic prediction value includes a longitudinal force prediction value and a lateral force prediction value; the adjustment on the fused tire force prediction value comprises: calculating a resultant force value of the preliminary fused tire force; comparing the resultant force value with a maximum allowable friction force calculated based on a current road adhesion coefficient and a vertical load; when the resultant force value is greater than the maximum allowable friction force, reducing the longitudinal force prediction value and the lateral force prediction value by a preset proportion until the resultant force value is equal to or less than the maximum allowable friction force.

8. A tire dynamics model modeling apparatus characterized by comprising: the device comprises: a first operation module configured to initialize a basic model of tire dynamics based on test data of dynamic performance of a tire, and determine a basic prediction value by using the initialized basic model; a second operation module configured to learn error distribution rules of the test data and the basic prediction value under complex working conditions by using a preset neural network model, and output an error compensation value corresponding to the basic prediction value; a fusion module configured to fuse the basic prediction value and the error compensation value by using a dynamic weighting fusion strategy to obtain a preliminary fused tire force; an adjustment module configured to correct the preliminary fused tire force for physical rationality to obtain a tire force meeting physical constraint conditions.

9. An electronic device, comprising: The electronic device comprises a memory and a processor, the memory stores a computer program, and the processor implements the tire dynamics modeling method of any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium storing a computer program, the computer program comprising instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1 to 9. The computer program is executed by the processor to implement the tire dynamics modeling method of any one of claims 1 to 7.

Images