Vehicle dynamics modeling method, system and device and computer readable storage medium

Abstract

A vehicle dynamics modeling method, system and device and a computer readable storage medium belong to the field of intelligent vehicle dynamics control, and comprise: obtaining a training data set, the training data set comprising a plurality of groups of training data, each group of training data comprising an input feature and a corresponding output feature, the input features comprise vehicle state quantities at multiple sampling moments before the current sampling moment and vehicle control quantities at the current sampling moment, and the output features comprise the vehicle state quantities at the current sampling moment; and performing model training according to the training data set to obtain a vehicle dynamics model so as to predict the output features when the input features are received. The vehicle dynamics model obtained based on training of the training data set can accurately predict the state quantity generated at the current moment in combination with the state quantities at multiple previous moments when receiving the control quantity, and the model can more accurately simulate the high time sequence correlation of the vehicle dynamics system and accurately restore the vehicle dynamics system.

Description

Technical Field

[0001] The present application relates to the field of intelligent vehicle dynamics control, and in particular to a vehicle dynamics modeling method, system, device and computer-readable storage medium. Background Art

[0002] In intelligent driving simulation tests, the vehicle dynamics model is an important part, and its accuracy will affect the credibility of the simulation test results. The traditional physical modeling method establishes differential equations by analyzing the lateral and longitudinal forces of the car, but the vehicle dynamics system is complex and uncertain, which makes physical modeling more difficult. Usually, the model is simplified, but the simplified dynamic model cannot reflect the essential characteristics of its dynamic system.

[0003] Although some mature commercial modeling software currently available has a high degree of physical restoration, it has a high price cost on the one hand. On the other hand, actual vehicles are equipped with complex actuators such as intelligent integrated braking systems (English: Integrated Power Brake; abbreviation: IPB) and intelligent driving algorithms, and it is difficult for commercial software to simulate the characteristics of actuators.

[0004] The data-driven model based on deep learning can perform high-precision fitting of input and output by learning a large amount of vehicle dynamics data collected from real vehicles / simulations, thereby describing the nonlinear and high-order dynamic effects of the vehicle and achieving higher dynamic accuracy. For example, the Chinese invention patent with application number CN202010643680.4 discloses a vehicle dynamics data-driven modeling method, which includes the following steps: collecting state quantity information of the vehicle to be modeled; designing a Deep EDMD network structure, using the Deep EDMD network to approximate the Koopman operator, linearly approximating the strongly coupled nonlinear vehicle dynamics model, and obtaining the experimental result data of the vehicle to be modeled; comparing the collected state quantity information with the obtained experimental result data, and drawing a comparison chart of the experimental results. However, the patent solution does not fully consider the high temporal correlation of vehicle dynamics samples, and cannot capture and use past information to predict future states. Summary of the invention

[0005] The present application provides a vehicle dynamics modeling method, system, device and computer-readable storage medium, which can solve the technical problem that vehicle dynamics modeling in the prior art does not fully consider the high temporal correlation of vehicle dynamics samples, resulting in the inability to accurately simulate the vehicle dynamics system.

[0006] In a first aspect, an embodiment of the present application provides a vehicle dynamics modeling method, the vehicle dynamics modeling method comprising: Acquire a training data set, the training data set comprising multiple sets of training data, each set of training data comprising input features and corresponding output features, the input features comprising vehicle state quantities at multiple sampling moments before the current sampling moment, and vehicle control quantities at the current sampling moment, and the output features comprising the vehicle state quantity at the current sampling moment; Model training is performed according to the training data set to obtain a vehicle dynamics model so as to predict output features when input features are received.

[0007] In combination with the first aspect, in one implementation, obtaining the training data set specifically includes the following steps: The training data set is obtained through real vehicle testing and simulation testing.

[0008] In combination with the first aspect, in one implementation, when performing the actual measurement test and the simulation test, a control instruction is obtained according to the expected state quantity output by the intelligent driving algorithm, and the control instruction includes a vehicle control quantity; The control instructions are sent to the real vehicle and the virtual vehicle respectively to control the operation of the real vehicle and the virtual vehicle and collect the vehicle status quantities.

[0009] In combination with the first aspect, in one implementation, the training data set includes training data of the vehicle under a plurality of typical operating conditions; The plurality of typical operating conditions include a coasting test, a longitudinal acceleration test, a braking test, and a handling stability test.

[0010] In combination with the first aspect, in one implementation, the vehicle control variable includes at least one of a motor torque, a brake master cylinder pressure, and a steering wheel steering angle; The vehicle state quantity includes at least one of a longitudinal vehicle speed, a lateral vehicle speed, a longitudinal acceleration, a lateral acceleration, a yaw angle, and a yaw angular velocity.

[0011] In combination with the first aspect, in one implementation, performing model training according to the training data set to obtain the vehicle dynamics model specifically includes the following steps: An encoder-decoder (ED) architecture based on a long short-term memory network (LSTM) is adopted, and the input features and output features are respectively used as the input and output of the architecture for training to obtain the vehicle dynamics model.

[0012] In combination with the first aspect, in one implementation, the method further includes: A validation dataset is constructed based on the training data collected from actual vehicle tests to verify and optimize the vehicle dynamics model.

[0013] In a second aspect, an embodiment of the present application provides a vehicle dynamics modeling system, the vehicle dynamics modeling system comprising: A data acquisition module, which is used to obtain a training data set, wherein the training data set includes multiple sets of training data, each set of training data includes input features and corresponding output features, wherein the input features include vehicle state quantities at multiple sampling moments before the current sampling moment, and vehicle control quantities at the current sampling moment, and the output features include the vehicle state quantity at the current sampling moment; A model training module is used to perform model training according to the training data set to obtain a vehicle dynamics model so as to predict output features when receiving input features.

[0014] In a third aspect, an embodiment of the present application provides a vehicle dynamics modeling device, which includes a processor, a memory, and a vehicle dynamics modeling program stored in the memory and executable by the processor, wherein when the vehicle dynamics modeling program is executed by the processor, the steps of the vehicle dynamics modeling method are implemented.

[0015] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, on which a vehicle dynamics modeling program is stored, wherein when the vehicle dynamics modeling program is executed by a processor, the steps of the vehicle dynamics modeling method are implemented.

[0016] The beneficial effects brought by the technical solution provided in the embodiments of the present application include: The present application constructs a training data set through the control quantity at the current moment, the state quantity at multiple moments before the current moment, and the state quantity at the current moment. The vehicle dynamics model trained according to the training data set can accurately predict the state quantity that will be generated at the current moment when receiving the control quantity in combination with the state quantities at multiple previous moments. The model can more accurately simulate the high temporal correlation of the vehicle dynamics system and accurately restore the vehicle dynamics system. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A schematic diagram of a flow chart of an embodiment of a vehicle dynamics modeling method of the present application; Figure 2 A schematic diagram of the architecture of an embodiment of a vehicle dynamics modeling system of the present application; Figure 3 This is a schematic diagram of the hardware structure of the vehicle dynamics modeling device involved in the embodiment of the present application. DETAILED DESCRIPTION

[0018] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0019] In order to make the objectives, technical solutions and advantages of the present application clearer, the implementation methods of the present application will be further described in detail below with reference to the accompanying drawings.

[0020] In a first aspect, an embodiment of the present application provides a vehicle dynamics modeling method.

[0021] In one embodiment, referring to Figure 1 , Figure 1 This is a flow chart of the first embodiment of the vehicle dynamics modeling method of this application. Figure 1 As shown, the vehicle dynamics modeling method includes: Step S1, obtaining a training data set, the training data set includes multiple sets of training data, each set of training data includes input features and corresponding output features, the input features include vehicle state quantities at multiple sampling moments before the current sampling moment, and vehicle control quantities at the current sampling moment, and the output features include the vehicle state quantity at the current sampling moment; Step S2: Perform model training according to the training data set to obtain a vehicle dynamics model to predict output features when receiving input features.

[0022] In this embodiment, the vehicle dynamics system is a time-varying system, and the output state quantity is usually closely related to the historical state quantity. Under the action of the input control quantity, the vehicle will generate a new dynamic state quantity, which will be passed to the next control cycle for the next optimization solution, thereby continuously solving and iterating to achieve continuous control of the vehicle.

[0023] Based on the above analysis, the present application constructs a training data set through the control quantity at the current moment, the state quantity at multiple moments before the current moment, and the state quantity at the current moment. The vehicle dynamics model trained according to the training data set can accurately predict the state quantity that will be generated at the current moment when receiving the control quantity in combination with the state quantities at multiple previous moments. The model can more accurately simulate the high temporal correlation of the vehicle dynamics system and accurately restore the vehicle dynamics system.

[0024] Furthermore, in one embodiment, the above-mentioned obtaining of the training data set specifically includes the following steps: The above training data set is obtained through real vehicle testing and simulation testing.

[0025] In this embodiment, before the vehicle dynamics model is constructed and trained, it is necessary to first determine the input and output data of the vehicle dynamics system and perform corresponding data collection. The real vehicle road data can truly restore the dynamic characteristics of the real vehicle, but the real vehicle data under extreme working conditions is difficult to obtain. Therefore, it is generally possible to supplement the input and output data of the vehicle dynamics system under extreme working conditions in a simulation environment. In the simulation environment and the actual environment, control signals are input to the vehicle dynamics system respectively to generate corresponding vehicle responses, and then the required input and output data are collected for the subsequent training of the vehicle dynamics model, so that the trained vehicle dynamics model can accurately predict the vehicle dynamics characteristics under various working conditions.

[0026] Further, in one embodiment, when performing the above-mentioned actual measurement test and the above-mentioned simulation test, the control instruction is obtained according to the expected state quantity output by the intelligent driving algorithm, and the above-mentioned control instruction includes the vehicle control quantity; The control instructions are sent to the real vehicle and the virtual vehicle respectively to control the operation of the real vehicle and the virtual vehicle and collect the vehicle status quantities.

[0027] In this embodiment, in the intelligent driving simulation test, the intelligent driving algorithm generally outputs the expected acceleration / steering angle, and after being processed by the actuator, the control instruction is input to the vehicle dynamics model, and then the corresponding state quantity is generated. In order to describe the relationship between this input and output, a data-driven method can be used to construct a neural network model. Correspondingly, it is necessary to process the expected state quantity output by the intelligent driving algorithm to obtain the control instruction, and send the control instruction to the real car and the virtual car respectively. The vehicle dynamics system of the real car and the virtual car adjusts its own state quantity according to the control instruction. In this process, the input control quantity and the output state quantity are collected to construct a training data set for subsequent neural network training. The vehicle dynamics model trained according to the training data set can accurately reflect the input and output relationship of the vehicle dynamics system, so that the vehicle dynamics system can be accurately simulated and restored.

[0028] Further, in one embodiment, the training data set includes training data of the vehicle under multiple typical operating conditions; The above-mentioned multiple typical working conditions include coasting test, longitudinal acceleration test, braking test, and handling stability test.

[0029] In this embodiment, multiple groups of typical working conditions are set according to national standard test conditions or ISO standard test conditions, such as coasting test, longitudinal acceleration test, braking test, and handling stability test. In different working condition tests, it is necessary to control the vehicle to travel at different speeds, accelerations, and steering angles based on the vehicle dynamics system, collect input control quantities and output state quantities during the vehicle driving process to construct a training data set, so that the vehicle dynamics model trained based on the training data set can accurately reflect the working state and performance of the vehicle dynamics system under different working conditions.

[0030] Further, in one embodiment, the vehicle control variable includes at least one of a motor torque, a brake master cylinder pressure, and a steering wheel steering angle; The vehicle state quantity includes at least one of a longitudinal vehicle speed, a lateral vehicle speed, a longitudinal acceleration, a lateral acceleration, a yaw angle, and a yaw angular velocity.

[0031] In this embodiment, for the simulation environment, three simulation softwares, Carsim, Simulink and VTD, are used for joint simulation. The input control quantity can be selected as motor torque, brake master cylinder pressure and steering wheel steering angle, and the output state quantity can be selected as longitudinal vehicle speed, lateral vehicle speed, longitudinal acceleration, lateral acceleration, yaw angle and yaw angular velocity. In Carsim, the initial speed of the vehicle is set to 0, and control instructions for acceleration, deceleration and steering are randomly given. The data acquisition length is 100s and the sampling frequency is 0.01s.

[0032] For the real vehicle road data, the data to be extracted is the same as above, and the data length is processed to 100s and the frequency is 0.01s.

[0033] Furthermore, in one embodiment, the above-mentioned model training is performed according to the training data set to obtain the vehicle dynamics model, which specifically includes the following steps: An encoder-decoder (ED) architecture based on a long short-term memory network (LSTM) is adopted, and the input features and output features are respectively used as the input and output of the architecture for training to obtain the above-mentioned vehicle dynamics model.

[0034] In this embodiment, since the changes in the dynamic state of the vehicle have strong temporal correlation, the structure of the neural network adopts the encoder-decoder ED architecture based on the long short-term memory network LSTM, that is, the ED-LSTM architecture, which can more accurately simulate and restore the strong temporal characteristics of the vehicle dynamics system and is more suitable for solving the task of predicting vehicle dynamics changes.

[0035] Furthermore, in one embodiment, the above method further includes: Step S3: construct a verification data set based on the training data collected from the actual vehicle test to verify and optimize the vehicle dynamics model.

[0036] In this embodiment, in the vehicle dynamics model verification at the vehicle level, the actual vehicle test data is used as reference data, the already built vehicle dynamics model is used for simulation testing, and the obtained simulation data is compared with the actual vehicle test data to analyze the consistency of the two. In practice, under the same vehicle driving action input, the vehicle state information output by the vehicle dynamics model verification and the actual vehicle is compared. Generally, the consistency between the two is required to reach more than 85%, and the prediction accuracy of the vehicle dynamics model is considered to meet the expectations. If it does not meet the expectations, the model is iteratively optimized according to the actual vehicle test data. For the actual vehicle data, CANoe can be used to establish a communication access programming language (English: Communication Access Programming Language; abbreviation: CAPL) and connect to the controller area network (English: Controller Area Network; abbreviation: CAN port) on the actual vehicle, send driving commands to the actual vehicle, and read and save CAN data.

[0037] In a specific embodiment, the input quantity of each data in the data set is the current time T, the motor torque, the brake master cylinder pressure, the steering wheel steering angle at time T, and the longitudinal vehicle speed, the lateral vehicle speed, the longitudinal acceleration, the lateral acceleration, the yaw angle, and the yaw angular velocity at time T-1, time T-2, and time T-3. The output is the longitudinal vehicle speed, the lateral vehicle speed, the longitudinal acceleration, the lateral acceleration, the yaw angle, and the yaw angular velocity at time T. That is, the state quantity that should be generated at the current moment is predicted through the control quantity at the current moment and the state quantity at the previous three moments.

[0038] A neural network model is built to learn the above data. The input parameters in the neural network model correspond to the control quantity at the current moment and the state quantity at the previous three moments, and the output parameters are the six output state quantities at the current moment. Combining the state quantities at multiple moments before the current moment with the control quantity at the current moment can more accurately predict the control quantity at the current moment, and can also more accurately simulate the vehicle dynamics system.

[0039] In summary, the present invention conducts data-driven vehicle dynamics modeling from three aspects: data collection, model building and training, and model accuracy and verification. It fully exploits the dynamic information contained in the real vehicle road data. For complex black box models, neural networks can better learn the relationship between input and output. Finally, a method for model accuracy verification is given. The vehicle dynamics model built in this paper has a high degree of restoration, good approximation effect, and high adaptability.

[0040] In a second aspect, an embodiment of the present application also provides a vehicle dynamics modeling system.

[0041] In one embodiment, referring to Figure 2 , Figure 2 This is a schematic diagram of the functional modules of an embodiment of the vehicle dynamics modeling device of the present application. Figure 2 As shown, the vehicle dynamics modeling system includes: A data acquisition module 1 is used to obtain a training data set, wherein the training data set includes multiple sets of training data, each set of training data includes input features and corresponding output features, wherein the input features include vehicle state quantities at multiple sampling moments before the current sampling moment, and vehicle control quantities at the current sampling moment, and the output features include the vehicle state quantity at the current sampling moment; The model training module 2 is used to perform model training according to the above training data set to obtain a vehicle dynamics model so as to predict output features when receiving input features.

[0042] In this embodiment, the vehicle dynamics system is a time-varying system, and the output state quantity is usually closely related to the historical state quantity. Under the action of the input control quantity, the vehicle will generate a new dynamic state quantity, which will be passed to the next control cycle for the next optimization solution, thereby continuously solving and iterating to achieve continuous control of the vehicle.

[0043] Based on the above analysis, the present application constructs a training data set through the control quantity at the current moment, the state quantity at multiple moments before the current moment, and the state quantity at the current moment. The vehicle dynamics model trained according to the training data set can accurately predict the state quantity that will be generated at the current moment when receiving the control quantity in combination with the state quantities at multiple previous moments. The model can more accurately simulate the high temporal correlation of the vehicle dynamics system and accurately restore the vehicle dynamics system.

[0044] Furthermore, in one embodiment, when the data acquisition module 1 obtains the training data set, the training data set is obtained through actual vehicle testing and simulation testing.

[0045] In this embodiment, before the vehicle dynamics model is constructed and trained, it is necessary to first determine the input and output data of the vehicle dynamics system and perform corresponding data collection. The real vehicle road data can truly restore the dynamic characteristics of the real vehicle, but the real vehicle data under extreme working conditions is difficult to obtain. Therefore, it is generally possible to supplement the input and output data of the vehicle dynamics system under extreme working conditions in a simulation environment. In the simulation environment and the actual environment, control signals are input to the vehicle dynamics system respectively to generate corresponding vehicle responses, and then the required input and output data are collected for the subsequent training of the vehicle dynamics model, so that the trained vehicle dynamics model can accurately predict the vehicle dynamics characteristics under various working conditions.

[0046] Further, in one embodiment, when performing the above-mentioned actual test and the above-mentioned simulation test, the data acquisition module 1 processes the desired state quantity output by the intelligent driving algorithm to obtain the control instruction, and the above-mentioned control instruction includes the vehicle control quantity; The control instructions are sent to the real vehicle and the virtual vehicle respectively to control the operation of the real vehicle and the virtual vehicle and collect the vehicle status quantities.

[0047] In this embodiment, in the intelligent driving simulation test, the intelligent driving algorithm generally outputs the expected acceleration / steering angle, and after being processed by the actuator, the control instruction is input to the vehicle dynamics model, and then the corresponding state quantity is generated. In order to describe the relationship between this input and output, a data-driven method can be used to construct a neural network model. Correspondingly, it is necessary to process the expected state quantity output by the intelligent driving algorithm to obtain the control instruction, and send the control instruction to the real car and the virtual car respectively. The vehicle dynamics system of the real car and the virtual car adjusts its own state quantity according to the control instruction. In this process, the input control quantity and the output state quantity are collected to construct a training data set for subsequent neural network training. The vehicle dynamics model trained according to the training data set can accurately reflect the input and output relationship of the vehicle dynamics system, so that the vehicle dynamics system can be accurately simulated and restored.

[0048] Further, in one embodiment, the training data set includes training data of the vehicle under multiple typical operating conditions; The above-mentioned multiple typical working conditions include coasting test, longitudinal acceleration test, braking test, and handling stability test.

[0049] In this embodiment, multiple groups of typical working conditions are set according to national standard test conditions or ISO standard test conditions, such as coasting test, longitudinal acceleration test, braking test, and handling stability test. In different working condition tests, it is necessary to control the vehicle to travel at different speeds, accelerations, and steering angles based on the vehicle dynamics system, collect input control quantities and output state quantities during the vehicle driving process to construct a training data set, so that the vehicle dynamics model trained based on the training data set can accurately reflect the working state and performance of the vehicle dynamics system under different working conditions.

[0050] Further, in one embodiment, the vehicle control variable includes motor torque, brake master cylinder pressure, and steering wheel steering angle; The vehicle state quantities include longitudinal vehicle speed, lateral vehicle speed, longitudinal acceleration, lateral acceleration, yaw angle, and yaw angular velocity.

[0051] In this embodiment, for the simulation environment, three simulation softwares, Carsim, Simulink and VTD, are used for joint simulation. The input control quantity can be selected as motor torque, brake master cylinder pressure and steering wheel steering angle, and the output state quantity can be selected as longitudinal vehicle speed, lateral vehicle speed, longitudinal acceleration, lateral acceleration, yaw angle and yaw angular velocity. In Carsim, the initial speed of the vehicle is set to 0, and control instructions for acceleration, deceleration and steering are randomly given. The data acquisition length is 100s and the sampling frequency is 0.01s.

[0052] For the real vehicle road data, the data to be extracted is the same as above, and the data length is processed to 100s and the frequency is 0.01s.

[0053] Furthermore, in one embodiment, the model training module 2 performs model training according to the training data set. When the vehicle dynamics model is obtained, based on the ED-LSTM architecture, the input features and output features are respectively trained as the input and output of the architecture to obtain the above-mentioned vehicle dynamics model.

[0054] In this embodiment, since the changes in the dynamic state of the vehicle have strong temporal correlation, the structure of the neural network adopts the encoder-decoder ED architecture based on the long short-term memory network LSTM, that is, the ED-LSTM architecture, which can more accurately simulate and restore the strong temporal characteristics of the vehicle dynamics system and is more suitable for solving the task of predicting vehicle dynamics changes.

[0055] Furthermore, in one embodiment, the model training module 2 is also used to construct a verification data set based on the training data collected from the actual vehicle test, so as to perform model verification and optimization on the vehicle dynamics model.

[0056] In this embodiment, in the vehicle dynamics model verification at the vehicle level, the actual vehicle test data is used as reference data, the established vehicle dynamics model is used for simulation testing, and the obtained simulation data is compared with the actual vehicle test data to analyze the consistency between the two. In practice, under the same vehicle driving action input, the vehicle state information output by the vehicle dynamics model verification and the actual vehicle is compared. Generally, the consistency between the two is required to reach more than 85%, and the prediction accuracy of the vehicle dynamics model is considered to meet the expectations. If it does not meet the expectations, the model is iteratively optimized according to the actual vehicle test data.

[0057] Among them, the functional implementation of each module in the above-mentioned vehicle dynamics modeling system corresponds to each step in the above-mentioned vehicle dynamics modeling method embodiment, and its functions and implementation processes are no longer repeated here.

[0058] In a third aspect, an embodiment of the present application provides a vehicle dynamics modeling device, which may be a personal computer (PC), a laptop computer, a server, or other device with data processing capabilities.

[0059] Reference Figure 3 , Figure 3 The hardware structure diagram of the vehicle dynamics modeling device involved in the embodiment of the present application is shown in FIG. In the embodiment of the present application, the vehicle dynamics modeling device may include a processor, a memory, a communication interface, and a communication bus.

[0060] The communication bus may be of any type and is used to interconnect the processor, the memory, and the communication interface.

[0061] The communication interface includes input / output (I / O) interface, physical interface and logical interface, etc., which are used to realize the interconnection of devices inside the vehicle dynamics modeling device, and the interface used to realize the interconnection between the vehicle dynamics modeling device and other devices (such as other computing devices or user devices). The physical interface can be an Ethernet interface, a fiber optic interface, an ATM interface, etc.; the user device can be a display, a keyboard, etc.

[0062] The memory can be various types of storage media, such as random access memory (RAM), read-only memory (ROM), non-volatile RAM (NVRAM), flash memory, optical storage, hard disk, programmable ROM (PROM), erasable PROM (EPROM), electrically erasable PROM (EEPROM), etc.

[0063] The processor may be a general-purpose processor, which may call the vehicle dynamics modeling program stored in the memory and execute the vehicle dynamics modeling method provided in the embodiment of the present application. For example, the general-purpose processor may be a central processing unit (CPU). The method executed when the vehicle dynamics modeling program is called may refer to the various embodiments of the vehicle dynamics modeling method of the present application, and will not be described in detail here.

[0064] Those skilled in the art will understand that Figure 3The hardware structure shown in the figure does not constitute a limitation on the present application, and may include more or less components than shown in the figure, or combine certain components, or arrange the components differently.

[0065] In a fourth aspect, an embodiment of the present application also provides a computer-readable storage medium.

[0066] The computer-readable storage medium of the present application stores a vehicle dynamics modeling program, wherein when the vehicle dynamics modeling program is executed by a processor, the steps of the vehicle dynamics modeling method as described above are implemented.

[0067] Among them, the method implemented when the vehicle dynamics modeling program is executed can refer to the various embodiments of the vehicle dynamics modeling method of the present application, and will not be repeated here.

[0068] It should be noted that the serial numbers of the above-mentioned embodiments of the present application are only for description and do not represent the advantages or disadvantages of the embodiments.

[0069] The terms "including" and "having" and any variations thereof in the specification and claims of this application and the above-mentioned drawings are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally includes steps or units that are not listed, or optionally includes other steps or units inherent to these processes, methods, products or devices. The terms "first", "second" and "third" are used to distinguish different objects, etc., and do not represent a sequence, nor do they limit "first", "second" and "third" to different types.

[0070] In the description of the embodiments of the present application, "exemplary", "for example" or "for example" are used to indicate examples, illustrations or descriptions. Any embodiment or design described as "exemplary", "for example" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary", "for example" or "for example" is intended to present related concepts in a specific way.

[0071] In the description of the embodiments of the present application, unless otherwise specified, “ / ” means or, for example, A / B can mean A or B; the “and / or” in the text is merely a description of 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. In addition, in the description of the embodiments of the present application, “multiple” refers to two or more than two.

[0072] In some processes described in the embodiments of the present application, multiple operations or steps that appear in a specific order are included, but it should be understood that these operations or steps may not be executed in the order in which they appear in the embodiments of the present application or in parallel, and the sequence number of the operation is only used to distinguish the different operations, and the sequence number itself does not represent any execution order. In addition, these processes may include more or fewer operations, and these operations or steps may be executed in sequence or in parallel, and these operations or steps may be combined.

[0073] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, disk, CD) as described above, and includes a number of instructions for a terminal device to execute the methods described in each embodiment of the present application.

[0074] The above are only preferred embodiments of the present application, and are not intended to limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A vehicle dynamics modeling method, characterized in that: The vehicle dynamics modeling method comprises: Acquire a training data set, the training data set comprising multiple sets of training data, each set of training data comprising input features and corresponding output features, the input features comprising vehicle state quantities at multiple sampling moments before the current sampling moment, and vehicle control quantities at the current sampling moment, and the output features comprising the vehicle state quantity at the current sampling moment; Model training is performed according to the training data set to obtain a vehicle dynamics model so as to predict output features when input features are received.

2. The vehicle dynamics modeling method according to claim 1, characterized in that: The step of obtaining the training data set specifically includes the following steps: The training data set is obtained through real vehicle testing and simulation testing.

3. The vehicle dynamics modeling method according to claim 2, characterized in that: When performing the actual test and the simulation test, the control instruction is obtained according to the expected state quantity output by the intelligent driving algorithm, and the control instruction includes the vehicle control quantity; The control instructions are sent to the real vehicle and the virtual vehicle respectively to control the operation of the real vehicle and the virtual vehicle and collect the vehicle status quantities.

4. The vehicle dynamics modeling method according to claim 1, characterized in that: The training data set includes training data of the vehicle under multiple typical working conditions; The plurality of typical operating conditions include a coasting test, a longitudinal acceleration test, a braking test, and a handling stability test.

5. The vehicle dynamics modeling method according to claim 1, characterized in that: The vehicle control variable includes at least one of a motor torque, a brake master cylinder pressure, and a steering wheel steering angle; The vehicle state quantity includes at least one of a longitudinal vehicle speed, a lateral vehicle speed, a longitudinal acceleration, a lateral acceleration, a yaw angle, and a yaw angular velocity.

6. The vehicle dynamics modeling method according to claim 1, characterized in that: The model training is performed according to the training data set to obtain the vehicle dynamics model, which specifically includes the following steps: An encoder-decoder (ED) architecture based on a long short-term memory network (LSTM) is adopted, and the input features and the output features are respectively used as the input and output of the architecture for training to obtain the vehicle dynamics model.

7. The vehicle dynamics modeling method according to claim 2, characterized in that: The method further comprises: A validation dataset is constructed based on the training data collected from actual vehicle tests to verify and optimize the vehicle dynamics model.

8. A vehicle dynamics modeling system, characterized in that: The vehicle dynamics modeling system comprises: A data acquisition module, which is used to obtain a training data set, wherein the training data set includes multiple sets of training data, each set of training data includes input features and corresponding output features, wherein the input features include vehicle state quantities at multiple sampling moments before the current sampling moment, and vehicle control quantities at the current sampling moment, and the output features include the vehicle state quantity at the current sampling moment; A model training module is used to perform model training according to the training data set to obtain a vehicle dynamics model so as to predict output features when receiving input features.

9. A vehicle dynamics modeling device, characterized in that: The vehicle dynamics modeling device includes a processor, a memory, and a vehicle dynamics modeling program stored in the memory and executable by the processor, wherein when the vehicle dynamics modeling program is executed by the processor, the steps of the vehicle dynamics modeling method as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a vehicle dynamics modeling program, wherein when the vehicle dynamics modeling program is executed by a processor, the steps of the vehicle dynamics modeling method according to any one of claims 1 to 7 are implemented.

Images