Neural Network-Based Vehicle Dynamics Modeling Methods, Devices, and Electronic Equipment

Abstract

This disclosure relates to a vehicle dynamics modeling method, apparatus, and electronic device based on a neural network. The method includes: acquiring real-vehicle test data for the current frame and corresponding historical set number of frames when the anti-lock braking system (ABS) is triggered, wherein the real-vehicle test data includes vehicle state variables and vehicle control variables; training the neural network using the control variables and vehicle state variables of the current frame and corresponding historical set number of frames to generate a neural network model with latency characteristics; and providing the neural network model to the vehicle's hardware-in-the-loop system as a real-time vehicle dynamics model. Therefore, in the ABS scenario, a vehicle dynamics model is constructed based on a neural network, simplifying and efficiing the process of establishing, debugging, and testing the vehicle dynamics model, and shortening the modeling cycle.

Description

Technical Field

[0001] This disclosure relates to the field of vehicle dynamics modeling technology, and in particular to a vehicle dynamics modeling method, apparatus and electronic device based on neural networks. Background Technology

[0002] During vehicle operation, vehicle dynamics models can reflect the vehicle's state. In actual development, modeling can be used for debugging. Related technologies utilize real-time simulation components (e.g., the CarSim real-time module) as a real-time operating system to run vehicle dynamics modules, enabling modeling of the entire vehicle's powertrain. This has wide applications in the automotive and construction machinery sectors, including drive-by-wire chassis and hardware-in-the-loop testing in autonomous driving.

[0003] However, the vehicle dynamics models of components in real-time simulations are basically based on measured data, requiring repeated parameter tuning and calibration before being compiled into the target real-time system. The whole process is quite cumbersome, and the parameter tuning and calibration processes are time-consuming and inefficient. Summary of the Invention

[0004] This disclosure provides a vehicle dynamics modeling method, apparatus, and electronic device based on neural networks.

[0005] According to a first aspect of the present disclosure, a vehicle dynamics modeling method based on a neural network is provided. The method includes: acquiring real vehicle test data of the current frame and a corresponding set number of historical frames when the anti-lock braking system is triggered, the real vehicle test data including vehicle state variables and vehicle control variables; training a neural network using the control variables and vehicle state variables of the current frame and the corresponding set number of historical frames to generate a neural network model with latency characteristics; and providing the neural network model to the vehicle's hardware-in-the-loop system as a real-time vehicle dynamics model.

[0006] In one embodiment of this disclosure, the step of obtaining real vehicle test data of the current frame and the corresponding historical set number of frames when the anti-lock braking system is triggered includes: determining different adhesion coefficients of the vehicle on the road surface; obtaining multiple initial test data of the current frame and the corresponding historical set number of frames when the anti-lock braking system is triggered under different adhesion coefficients; and filtering the multiple initial test data for abnormal data to obtain filtered real vehicle test data.

[0007] In one embodiment of this disclosure, the vehicle state quantities include at least one of the following: vehicle longitudinal velocity, vehicle lateral velocity, vehicle yaw rate, vehicle longitudinal acceleration, vehicle lateral acceleration, vehicle front left wheel velocity, vehicle front right wheel velocity, vehicle rear left wheel velocity, and vehicle rear right wheel velocity.

[0008] In one embodiment of this disclosure, the control quantities include at least one of the following: vehicle steering wheel angle, front drive torque of the wheels, rear drive torque of the wheels, front-circuit wheel braking pressure of the vehicle, and rear-circuit wheel braking pressure of the vehicle.

[0009] In one embodiment of this disclosure, training a neural network using the control quantity and vehicle state quantity of the current frame and the corresponding set number of historical frames to generate a neural network model with latency characteristics includes: determining a neural network, wherein the neural network is a fully connected feedforward neural network, and the amount of data of each neuron in the feedforward neural network is adjusted based on the vehicle state quantity and the vehicle control quantity; using the control quantity of the current frame and the corresponding set number of historical frames as the input of the feedforward neural network, and using the vehicle state quantity of the current frame and the corresponding set number of historical frames as the output of the feedforward neural network to train an initial neural network model with latency characteristics; and iteratively optimizing the initial neural network model using real vehicle test data of the current frame and the corresponding set number of historical frames in different time periods to obtain a neural network model that meets preset optimization conditions.

[0010] According to a second aspect of the present disclosure, a vehicle dynamics modeling device based on a neural network is also provided. The device includes: an acquisition module, configured to acquire real vehicle test data of the current frame and a corresponding set number of historical frames when the anti-lock braking system is triggered, the real vehicle test data including vehicle state variables and vehicle control variables; a training module, configured to train the neural network using the control variables and vehicle state variables of the current frame and the corresponding set number of historical frames to generate a neural network model with latency characteristics; and a retrieval module, configured to provide the neural network model to the vehicle's hardware-in-the-loop system as a real-time vehicle dynamics model.

[0011] In one embodiment of this disclosure, the acquisition module is specifically used to: determine different adhesion coefficients of the vehicle on the road surface; collect multiple initial test data of the current frame and corresponding historical set number of frames when the vehicle triggers the anti-lock braking system under different adhesion coefficients; and filter the multiple initial test data for abnormal data to obtain filtered real vehicle test data.

[0012] In one embodiment of this disclosure, the vehicle state quantities include at least one of the following: vehicle longitudinal velocity, vehicle lateral velocity, vehicle yaw rate, vehicle longitudinal acceleration, vehicle lateral acceleration, vehicle front left wheel velocity, vehicle front right wheel velocity, vehicle rear left wheel velocity, and vehicle rear right wheel velocity.

[0013] In one embodiment of this disclosure, the control quantities include at least one of the following: vehicle steering wheel angle, front drive torque of the wheels, rear drive torque of the wheels, front-circuit wheel braking pressure of the vehicle, and rear-circuit wheel braking pressure of the vehicle.

[0014] In one embodiment of this disclosure, the training module is specifically used for: determining a neural network, wherein the neural network is a fully connected feedforward neural network, and the amount of data of each neuron in the feedforward neural network is adjusted based on the vehicle state quantity and the vehicle control quantity; using the control quantity of the current frame and the corresponding set number of historical frames as the input of the feedforward neural network, and using the vehicle state quantity of the current frame and the corresponding set number of historical frames as the output of the feedforward neural network, to train an initial neural network model with delay characteristics; and iteratively optimizing the initial neural network model using real vehicle test data of the current frame and the corresponding set number of historical frames in different time periods to obtain a neural network model that meets preset optimization conditions.

[0015] According to a third aspect of the present disclosure, an electronic device is also provided, comprising: a processor; a memory for storing processor-executable instructions; wherein the processor is configured to: implement the steps of the neural network-based vehicle dynamics modeling method as described above.

[0016] According to a fourth aspect of the present disclosure, a non-transitory computer-readable storage medium is also provided, wherein when instructions in the storage medium are executed by a processor, the processor is able to execute the neural network-based vehicle dynamics modeling method as described above.

[0017] The technical solutions provided by the embodiments of this disclosure bring at least the following beneficial effects: This method acquires real-vehicle test data for the current frame and corresponding historical set frame numbers when the anti-lock braking system (ABS) is triggered. The real-vehicle test data includes vehicle state variables and vehicle control variables. Using the control variables and vehicle state variables from the current frame and corresponding historical set frame numbers, a neural network is trained to generate a neural network model with latency characteristics. This neural network model is then provided to the vehicle's hardware-in-the-loop system as a real-time vehicle dynamics model. Therefore, by constructing a vehicle dynamics model based on a neural network in the ABS scenario, the process of establishing, debugging, and testing the vehicle dynamics model is simplified and streamlined, and the modeling cycle is shortened.

[0018] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0019] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure, and are not intended to unduly limit this disclosure.

[0020] Figure 1 This is a flowchart illustrating a neural network-based vehicle dynamics modeling method according to an embodiment of this disclosure; Figure 2 This is a schematic diagram illustrating the principle of a neural network-based vehicle dynamics modeling method according to an embodiment of this disclosure; Figure 3 This is a schematic diagram of the structure of a neural network-based vehicle dynamics modeling device according to an embodiment of the present disclosure; Figure 4 This is a structural block diagram of an electronic device according to an exemplary embodiment of the present disclosure. Detailed Implementation

[0021] To enable those skilled in the art to better understand the technical solutions of this disclosure, the technical solutions in the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings.

[0022] It should be noted that the terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.

[0023] During vehicle operation, vehicle dynamics models can reflect the vehicle's state. In actual development, modeling can be used for debugging. Related technologies utilize real-time simulation components (e.g., the CarSim real-time module) as a real-time operating system to run vehicle dynamics modules, enabling modeling of the entire vehicle's powertrain. This has wide applications in the automotive and construction machinery sectors, including drive-by-wire chassis and hardware-in-the-loop testing in autonomous driving.

[0024] However, the vehicle dynamics models of components in real-time simulations are basically based on measured data, requiring repeated parameter tuning and calibration before being compiled into the target real-time system. The whole process is quite cumbersome, and the parameter tuning and calibration processes are time-consuming and inefficient.

[0025] Figure 1This is a flowchart illustrating a neural network-based vehicle dynamics modeling method according to an embodiment of the present disclosure. It should be noted that the neural network-based vehicle dynamics modeling method of this embodiment can be applied to a neural network-based vehicle dynamics modeling device, which can be configured in an electronic device to enable the electronic device to perform neural network-based vehicle dynamics modeling functions.

[0026] The electronic device can be any device with computing capabilities, such as a terminal device or a server. The following embodiments use a terminal device as an example for illustration.

[0027] like Figure 1 As shown, the method includes the following steps: Step 101: Obtain the real vehicle test data of the current frame and the corresponding historical set frame number when the anti-lock braking system is triggered. The real vehicle test data includes vehicle state variables and vehicle control variables.

[0028] In this embodiment of the disclosure, in order to improve the reliability and applicability of test data in the development of the Anti-lock Braking System (ABS), as an example, different adhesion coefficients of the vehicle on the road surface are determined; multiple initial test data of the current frame and the corresponding historical set number of frames when the vehicle triggers the ABS under different adhesion coefficients; and abnormal data are filtered from the multiple initial test data to obtain the filtered real vehicle test data.

[0029] Specifically, the different coefficients of adhesion for a vehicle on the road surface are related to the vehicle's speed. For example, on a high-adhesion road surface, the vehicle is accelerated from a standstill to 50 km / h, 60 km / h, 70 km / h, 80 km / h, 90 km / h, and 100 km / h. The same applies to a low-adhesion road surface. The maximum acceleration speed is adjusted appropriately, and then the anti-lock braking system (ABS) is triggered until the vehicle comes to a standstill again. Data from the entire process is recorded, and outlier data is filtered to obtain the filtered real-vehicle test data.

[0030] Among them, abnormal data screening can be done by selecting 10 sets of data for each speed range, while maintaining the uniformity of the data and ensuring that no bad data appears.

[0031] It should be noted that vehicle state variables include at least one of the following: vehicle longitudinal speed, vehicle lateral speed, vehicle yaw rate, vehicle longitudinal acceleration, vehicle lateral acceleration, vehicle front left wheel speed, vehicle front right wheel speed, vehicle rear left wheel speed, and vehicle rear right wheel speed. Control variables include at least one of the following: vehicle steering wheel angle, front drive torque of the wheels, rear drive torque of the wheels, braking pressure of the front-circuit wheels, and braking pressure of the rear-circuit wheels.

[0032] Step 102: Train the neural network using the control quantities and vehicle state quantities of the current frame and the corresponding historical set frame number to generate a neural network model with latency characteristics.

[0033] The historical frame rate can be set to 3 to 4 frames, but is not limited to this.

[0034] In this embodiment, due to the inherent delay in the response of braking, vehicle steering, and drive subsystems to continuous control inputs, the control inputs of 3 to 4 historical frames and vehicle state inputs are incorporated into the neural network input design. This allows the neural network to consider the influence of historical frames during optimization, resulting in a more reasonable neural network model with latency characteristics. As an example: the neural network is determined to be a fully connected feedforward neural network. The amount of data for each neuron in the feedforward neural network is adjusted based on the vehicle state and control inputs. The control inputs of the current frame and the corresponding set number of historical frames are used as the inputs to the feedforward neural network, and the vehicle state inputs of the current frame and the corresponding set number of historical frames are used as the outputs, training an initial neural network model with latency characteristics. Real-vehicle test data from different time periods (current frame and corresponding set number of historical frames) are used to iteratively optimize the initial neural network model to obtain a neural network model that meets preset optimization conditions.

[0035] For example, preset optimization conditions could be that the classification accuracy reaches a threshold or the training loss decreases to a stable value within a specified number of rounds, but these are not the only possibilities.

[0036] Step 103: The neural network model is provided to the vehicle's hardware-in-the-loop system as a real-time vehicle dynamics model.

[0037] In this embodiment of the disclosure, the optimized neural network model (e.g., saved as a .h5 file) needs to be deployed in real-time within a real-time operating system (e.g., NI Linux RT system). The neural network model will be further encapsulated using a software framework (e.g., NI Veristand Framework) embedded in the configuration of the real-time test application, and then further encapsulated using a programming language (e.g., C++) into a dynamic link library that can run on the real-time operating system (e.g., encapsulated via a .so file format). This exposes the input and output interfaces of the neural network model, and it can be called through hardware-in-the-loop testing tools (e.g., Veristand) or graphical programming environments (e.g., LabVIEW), providing access to the host computer and boards throughout the hardware-in-the-loop system.

[0038] In summary, the real-time vehicle dynamics model in ABS scenarios is closer to real vehicle data and better reflects the characteristics of real vehicles. It can reproduce, trace, and solve problems in a bench environment without the need for a real vehicle.

[0039] The vehicle dynamics modeling method based on neural networks disclosed in this embodiment acquires real-vehicle test data for the current frame and corresponding historical set number of frames when the anti-lock braking system (ABS) is triggered. The real-vehicle test data includes vehicle state variables and vehicle control variables. The neural network is trained using the control variables and vehicle state variables of the current frame and corresponding historical set number of frames to generate a neural network model with latency characteristics. This neural network model is then provided to the vehicle's hardware-in-the-loop system as a real-time vehicle dynamics model. Therefore, by constructing a vehicle dynamics model based on neural networks in the ABS scenario, the process of establishing, debugging, and testing the vehicle dynamics model is simplified and made more efficient, and the modeling cycle is shortened.

[0040] Figure 2 This is a schematic diagram illustrating the principle of a neural network-based vehicle dynamics modeling method according to an embodiment of this disclosure.

[0041] Specifically, the process involves acquiring real-vehicle test data and designing scenarios. For example, data-driven modeling of vehicle dynamics under ABS scenarios is performed, and data on the real vehicle's states at different adhesion coefficients is recorded. Initial test data is then filtered, removing unqualified data and striving for a more uniform data distribution to obtain accurate real-vehicle test data. The input and output of the neural network are designed based on the input and output requirements of the vehicle's own test bench, such as the vehicle's control variables and state variables. During the initial test data filtering, the neural network model's input and output can be arranged accordingly. A neural network framework is established, and the historical frame rate of input is adjusted for different continuous input and output values ​​to predict the delay caused by the inertia of the vehicle's own subsystems, thus training an initial neural network model with delay characteristics. This initial neural network model is iteratively refined until it reaches its optimal target and exhibits a certain degree of generalization ability (meeting preset optimization conditions), resulting in the final neural network model. This model is then encapsulated through standardized interfaces and tools (e.g., the NI Veristand C++ framework) to serve as a real-time vehicle dynamics model, capable of being loaded and run in a real-time operating system (NI LinuxRT). It can successfully call the vehicle dynamics model from the host computer and the board in the hardware-in-the-loop system.

[0042] Figure 3 This is a schematic diagram of the structure of a neural network-based vehicle dynamics modeling device according to an embodiment of the present disclosure.

[0043] like Figure 3As shown, the vehicle dynamics modeling device based on neural networks includes an acquisition module 301, a training module 302, and a retrieval module 303.

[0044] The acquisition module 301 is used to acquire real vehicle test data of the current frame and the corresponding set number of historical frames when the anti-lock braking system is triggered. The real vehicle test data includes vehicle state variables and vehicle control variables. The training module 302 is used to train the neural network using the control variables and vehicle state variables of the current frame and the corresponding set number of historical frames to generate a neural network model with latency characteristics. The retrieval module 303 is used to provide the neural network model to the vehicle's hardware-in-the-loop system as a real-time vehicle dynamics model.

[0045] In one embodiment of this disclosure, the acquisition module 301 is specifically used to: determine different adhesion coefficients of the vehicle on the road surface; collect multiple initial test data of the current frame and corresponding historical set number of frames when the vehicle triggers the anti-lock braking system under different adhesion coefficients; and filter the multiple initial test data for abnormal data to obtain filtered real vehicle test data.

[0046] In one embodiment of this disclosure, the vehicle state quantities include at least one of the following: vehicle longitudinal velocity, vehicle lateral velocity, vehicle yaw rate, vehicle longitudinal acceleration, vehicle lateral acceleration, vehicle front left wheel velocity, vehicle front right wheel velocity, vehicle rear left wheel velocity, and vehicle rear right wheel velocity.

[0047] In one embodiment of this disclosure, the control quantities include at least one of the following: vehicle steering wheel angle, front drive torque of the wheels, rear drive torque of the wheels, front-circuit wheel braking pressure of the vehicle, and rear-circuit wheel braking pressure of the vehicle.

[0048] In one embodiment of this disclosure, the training module 302 is specifically used for: determining a neural network, wherein the neural network is a fully connected feedforward neural network, and the amount of data of each neuron in the feedforward neural network is adjusted based on the vehicle state quantity and the vehicle control quantity; using the control quantity of the current frame and the corresponding set number of historical frames as the input of the feedforward neural network, and using the vehicle state quantity of the current frame and the corresponding set number of historical frames as the output of the feedforward neural network, to train an initial neural network model with delay characteristics; and iteratively optimizing the initial neural network model using real vehicle test data of the current frame and the corresponding set number of historical frames in different time periods to obtain a neural network model that meets preset optimization conditions.

[0049] The vehicle dynamics modeling device based on a neural network disclosed in this embodiment acquires real-vehicle test data for the current frame and corresponding historical set number of frames when the anti-lock braking system (ABS) is triggered. The real-vehicle test data includes vehicle state variables and vehicle control variables. The neural network is trained using the control variables and vehicle state variables of the current frame and corresponding historical set number of frames to generate a neural network model with latency characteristics. The neural network model is then provided to the vehicle's hardware-in-the-loop system as a real-time vehicle dynamics model. Therefore, in the ABS scenario, a vehicle dynamics model is constructed based on a neural network, simplifying and efficiing the process of establishing, debugging, and testing the vehicle dynamics model, and shortening the modeling cycle.

[0050] According to a third aspect of the present disclosure, an electronic device is also provided, comprising: a processor; and a memory for storing processor-executable instructions, wherein the processor is configured to: implement the neural network-based vehicle dynamics modeling method as described above.

[0051] To implement the above embodiments, this disclosure also proposes a storage medium.

[0052] When the instructions in the storage medium are executed by the processor, the processor is able to execute the neural network-based vehicle dynamics modeling method as described above.

[0053] To implement the above embodiments, this disclosure also provides a computer program product.

[0054] When the computer program product is executed by the processor of the electronic device, it enables the electronic device to perform the above-described method.

[0055] Figure 4 This is a structural block diagram of an electronic device according to an exemplary embodiment. Figure 4 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments disclosed herein.

[0056] like Figure 4 As shown, the electronic device 400 includes a processor 401, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 402 or a program loaded from memory 406 into a random access memory (RAM) 403. The RAM 403 also stores various programs and data required for the operation of the electronic device 400. The processor 401, ROM 402, and RAM 403 are interconnected via a bus 404. An input / output (I / O) interface 405 is also connected to the bus 404.

[0057] The following components are connected to I / O interface 405: memory 406 including hard disks, etc.; and communication section 407 including network interface cards such as local area network (LAN) cards, modems, etc., communication section 407 performs communication processing via a network such as the Internet; and driver 408 is also connected to I / O interface 405 as needed.

[0058] In particular, according to embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 407. When the computer program is executed by processor 401, it performs the functions defined in the methods of this disclosure.

[0059] In an exemplary embodiment, a storage medium including instructions is also provided, such as a memory including instructions, which can be executed by a processor 401 of an electronic device 400 to perform the above-described method. Optionally, the storage medium may be a non-transitory computer-readable storage medium, such as a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device.

[0060] In this disclosure, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in connection with an instruction execution system, apparatus, or device. In this disclosure, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, which can transmit, propagate, or transfer a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wireless, wireline, optical fiber, RF, etc., or any suitable combination thereof.

[0061] Furthermore, the term “exemplary” is used herein to mean serving as an example, instance, or illustration. Any aspect or design described herein as “exemplary” is not necessarily to be construed as advantageous compared to other aspects or designs. Rather, the use of the term “exemplary” is intended to present the concept in a concrete manner. As used herein, the term “or” is intended to mean an inclusive “or” rather than an exclusive “or.” That is, unless otherwise specified or clear from the context, “X applies A or B” is intended to mean any of the natural inclusive arrangements. That is, “X applies A or B” satisfies any of the foregoing instances if X applies A; X applies B; or both X applies A and B. Additionally, unless otherwise specified or clear from the context to refer to the singular form, the articles “a” and “an” as used in this application and the appended claims are generally understood to mean “one or more.”

[0062] Similarly, although this disclosure has been shown and described with respect to one or more implementations, equivalent variations and modifications will occur to those skilled in the art upon reading and understanding this specification and the accompanying drawings. This disclosure includes all such modifications and variations and is limited only by the scope of the claims. In particular, with respect to the various functions performed by the components described above (e.g., elements, resources, etc.), unless otherwise indicated, the terminology used to describe such components is intended to correspond to any component (functionally equivalent) that performs the specific function of the described component, even if structurally not equivalent to the disclosed structure. Furthermore, although specific features of this disclosure may have been disclosed with respect to only one of several implementations, such features may be combined with one or more other features of other implementations, as may be desired and advantageous to any given or particular application. Moreover, with regard to the terms “comprising,” “owning,” “having,” “having,” or variations thereof as used in the detailed description or claims, such terms are intended to be inclusive in a manner similar to the term “including.”

[0063] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This disclosure is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the following claims.

[0064] It should be understood that this disclosure is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this disclosure is limited only by the appended claims.

Claims

1. A vehicle dynamics modeling method based on neural networks, characterized in that, The method includes: Acquire real vehicle test data for the current frame and the corresponding historical set number of frames when the anti-lock braking system is triggered. The real vehicle test data includes vehicle state variables and vehicle control variables. The neural network is trained using the control quantities and vehicle state quantities of the current frame and the corresponding historical set frame number to generate a neural network model with latency characteristics. The neural network model is provided to the vehicle's hardware-in-the-loop system as a real-time vehicle dynamics model.

2. The method according to claim 1, characterized in that, The acquisition of real-vehicle test data for the current frame and corresponding historical set frame numbers when the anti-lock braking system is triggered includes: Determine the different coefficients of adhesion for vehicles on the road surface; The initial test data of the current frame and the corresponding historical set frame number when the vehicle triggers the anti-lock braking system under different adhesion coefficients; The multiple initial test data are filtered for abnormal data to obtain the filtered real vehicle test data.

3. The method according to claim 2, characterized in that, The vehicle state quantities include at least one of the following: vehicle longitudinal speed, vehicle lateral speed, vehicle yaw rate, vehicle longitudinal acceleration, vehicle lateral acceleration, vehicle front left wheel speed, vehicle front right wheel speed, vehicle rear left wheel speed, and vehicle rear right wheel speed.

4. The method according to claim 2, characterized in that, The control quantities include at least one of the following: vehicle steering wheel angle, front drive torque of the wheels, rear drive torque of the wheels, front-circuit wheel braking pressure of the vehicle, and rear-circuit wheel braking pressure of the vehicle.

5. The method according to claim 1, characterized in that, The step of training the neural network using the control quantities and vehicle state quantities of the current frame and the corresponding historical set number of frames to generate a neural network model with latency characteristics includes: The neural network is defined as a fully connected feedforward neural network, wherein the amount of data of each neuron in the feedforward neural network is adjusted based on the vehicle state variables and the vehicle control variables. The control quantities of the current frame and the corresponding set number of historical frames are used as the input of the feedforward neural network, and the vehicle state quantities of the current frame and the corresponding set number of historical frames are used as the output of the feedforward neural network to train an initial neural network model with delay characteristics. The initial neural network model is iteratively optimized using real vehicle test data of the current frame and the corresponding set number of historical frames at different time periods to obtain a neural network model that meets the preset optimization conditions.

6. A vehicle dynamics modeling device based on neural networks, characterized in that, The device includes: The acquisition module is used to acquire the real vehicle test data of the current frame and the corresponding historical set number of frames when the anti-lock braking system is triggered. The real vehicle test data includes vehicle state variables and vehicle control variables. The training module is used to train the neural network using the control quantities and vehicle state quantities of the current frame and the corresponding historical set number of frames, so as to generate a neural network model with latency characteristics. The retrieval module is used to provide the neural network model to the vehicle's hardware-in-the-loop system as a real-time vehicle dynamics model.

7. The apparatus according to claim 6, characterized in that, The acquisition module is specifically used for: Determine the different coefficients of adhesion for vehicles on the road surface; The initial test data of the current frame and the corresponding historical set frame number when the vehicle triggers the anti-lock braking system under different adhesion coefficients; The multiple initial test data are filtered for abnormal data to obtain the filtered real vehicle test data.

8. The apparatus according to claim 7, characterized in that, The vehicle state quantities include at least one of the following: vehicle longitudinal speed, vehicle lateral speed, vehicle yaw rate, vehicle longitudinal acceleration, vehicle lateral acceleration, vehicle front left wheel speed, vehicle front right wheel speed, vehicle rear left wheel speed, and vehicle rear right wheel speed.

9. The apparatus according to claim 7, characterized in that, The control quantities include at least one of the following: vehicle steering wheel angle, front drive torque of the wheels, rear drive torque of the wheels, front-circuit wheel braking pressure of the vehicle, and rear-circuit wheel braking pressure of the vehicle.

10. The apparatus according to claim 6, characterized in that, The training module is specifically used for: The neural network is defined as a fully connected feedforward neural network, wherein the amount of data of each neuron in the feedforward neural network is adjusted based on the vehicle state variables and the vehicle control variables. The control quantities of the current frame and the corresponding set number of historical frames are used as the input of the feedforward neural network, and the vehicle state quantities of the current frame and the corresponding set number of historical frames are used as the output of the feedforward neural network to train an initial neural network model with delay characteristics. The initial neural network model is iteratively optimized using real vehicle test data of the current frame and the corresponding set number of historical frames at different time periods to obtain a neural network model that meets the preset optimization conditions.

11. An electronic device, characterized in that, include: processor; Memory used to store the processor's executable instructions; The processor is configured as follows: The steps for implementing the neural network-based vehicle dynamics modeling method as described in any one of claims 1 to 5.

12. A non-transitory computer-readable storage medium, characterized in that, When the instructions in the storage medium are executed by the processor, the processor is able to perform the neural network-based vehicle dynamics modeling method as described in any one of claims 1 to 5.

Images