Method, device, controller, vehicle and program for calibrating a vehicle body stabilization system

Abstract

Embodiments of the present disclosure relate to a method, apparatus, controller, vehicle and computer program product for calibrating a vehicle body stability system. The method comprises obtaining vehicle model data of a vehicle. The method further comprises obtaining brake system data of a brake system of the vehicle. The method further comprises generating, according to the vehicle model data and the brake system data, a parameter for calibrating the vehicle body stability system using a generative neural network model. According to the method of the embodiments of the present disclosure, the parameter for calibrating the vehicle body stability system can be efficiently generated by using the generative neural network model based on the data related to the vehicle, the accuracy and efficiency of calibrating the vehicle body stability system are improved, and the cost of calibrating the vehicle body stability system is reduced.

Description

Technical Field

[0001] Embodiments of this disclosure relate to the field of vehicles, and more particularly to methods, apparatus, controllers, vehicles, and computer program products for calibrating the vehicle stability system of a vehicle. Background Technology

[0002] As vehicle models become increasingly diverse, so too do vehicle stability control systems (ESCs). ESCs are used to control the vehicle's braking system in emergency situations, enabling functions such as anti-lock braking and anti-skid braking to ensure vehicle safety. Typically, for new vehicles, the ESC is designed with a set of default parameters. However, during actual driving, the ESC under these default parameters may not accurately and stably control the braking system of the new vehicle. Therefore, the parameters of the ESC for new vehicles need to be calibrated or recalibrated to ensure driving safety. Because ESCs have many parameters, this calibration or recalibration usually requires highly experienced engineers. Summary of the Invention

[0003] Embodiments of this disclosure provide a method, apparatus, controller, vehicle, and computer program product for calibrating a vehicle's body stability system.

[0004] According to a first aspect of this disclosure, a method for calibrating a vehicle stability system is provided. The method includes acquiring vehicle model data. The method also includes acquiring braking system data of the vehicle's braking system. Furthermore, the method includes generating parameters for calibrating the vehicle stability system using a generative neural network model based on the vehicle model data and the braking system data.

[0005] According to a second aspect of this disclosure, an apparatus for calibrating a vehicle stability system is provided. The apparatus includes a first acquisition unit configured to acquire vehicle model data. The apparatus also includes a second acquisition unit configured to acquire braking system data of the vehicle's braking system. Furthermore, the apparatus includes a parameter generation unit configured to generate parameters for calibrating the vehicle stability system using a generative neural network model based on the vehicle model data and the braking system data.

[0006] According to a third aspect of this disclosure, a controller is provided. The controller includes at least one processor; and a memory coupled to the at least one processor and having instructions stored thereon, which, when executed by the at least one processor, cause the controller to implement the method according to a first aspect of this disclosure.

[0007] According to a fourth aspect of this disclosure, a vehicle is provided. The vehicle includes a controller according to a third aspect of this disclosure.

[0008] According to a fifth aspect of this disclosure, a computer program product is provided. The computer program product stores computer-executable instructions, which are executed by a processor to implement the method according to a first aspect of this disclosure.

[0009] In a sixth aspect of this disclosure, a computer-readable storage medium is provided. The computer-readable storage medium stores computer-executable instructions, which are executed by a processor to implement the method according to a first aspect of this disclosure. Attached Figure Description

[0010] The above and other objects, features and advantages of this disclosure will become more apparent from the accompanying drawings, in which like reference numerals generally denote like parts.

[0011] Figure 1 The illustration shows a schematic diagram of an example environment in which the apparatus and / or methods according to embodiments of the present disclosure may be implemented.

[0012] Figure 2 The illustration shows a flowchart of a method for calibrating a vehicle's vehicle stability system according to an embodiment of the present disclosure.

[0013] Figure 3 A flowchart illustrating the process of generating parameters according to an embodiment of the present disclosure is shown.

[0014] Figure 4 A flowchart illustrating the process of selecting parameters according to an embodiment of the present disclosure is shown.

[0015] Figure 5 The illustration shows an embodiment of the present disclosure. Figure 4 The diagram illustrates the process of using a generative neural network model to generate parameters.

[0016] Figure 6 A flowchart illustrating the process of selecting parameters according to another embodiment of the present disclosure is shown.

[0017] Figure 7 The illustration shows a flowchart of the process of training a generative neural network model according to an embodiment of the present disclosure.

[0018] Figure 8 The illustration shows a schematic block diagram of an apparatus for calibrating a vehicle's vehicle stability system according to an embodiment of the present disclosure.

[0019] Figure 9 A schematic block diagram of an example of an example device suitable for implementing embodiments of the present disclosure is illustrated.

[0020] In the various figures, the same or corresponding reference numerals indicate the same or corresponding parts. Detailed Implementation

[0021] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0022] In the description of embodiments of this disclosure, the term "comprising" and similar terms should be understood as open-ended inclusion, i.e., "including but not limited to". The term "based on" should be understood as "at least partially based on". The term "one embodiment" or "the embodiment" should be understood as "at least one embodiment". The terms "first", "second", etc., may refer to different or the same objects. Other explicit and implicit definitions may also be included below.

[0023] As mentioned earlier, due to the numerous parameters of a vehicle stability system (VSS), calibration or recalibration typically requires highly experienced engineers. Generally, after an engineer adjusts at least some of the VSS parameters, a real-vehicle test is conducted. During this test, the VSS is calibrated using the currently adjusted set of parameters, and the vehicle is driven in as many different environments as possible to obtain test data under as many conditions as possible (e.g., more than 150 conditions). This data includes curves (also called traces) of driving data such as speed, distance, and acceleration. The engineer then analyzes the shape of these curves, adjusts one or more parameters based on the analysis, and repeats the real-vehicle test. This process is repeated until all curves obtained from the real-vehicle test meet the engineer's requirements. Therefore, this method of calibrating (or recalibrating) a VSS consumes significant resources in terms of manpower, time, vehicle wear and tear, space, and extreme testing environments. It also heavily relies on the engineer's experience, leading to significant discrepancies in calibration between engineers and making it difficult to guarantee accuracy.

[0024] To address at least the aforementioned and other potential problems, embodiments of this disclosure provide a method for calibrating a vehicle's stability control system. The method includes acquiring vehicle model data. The method also includes acquiring braking system data of the vehicle's braking system. Furthermore, the method includes using a generative neural network model to generate parameters for calibrating the stability control system based on the vehicle model data and the braking system data. The method according to embodiments of this disclosure enables the efficient and automatic generation of parameters for calibrating the vehicle's stability control system using a generative neural network model based on vehicle-related data, improving the accuracy and efficiency of vehicle stability control system calibration and reducing the cost of calibrating the stability control system.

[0025] The embodiments of this disclosure will now be described in further detail with reference to the accompanying drawings. Figure 1 The illustration shows a schematic diagram of an example environment 100 in which the devices and / or methods according to embodiments of the present disclosure may be implemented. Figure 1 Example environment 100 illustrates road 110 and vehicles 120, 130, 140, and 150 traveling on the road. Any of vehicles 120, 130, 140, and 150 may have a vehicle stability system, such as Electronic Stability Program (ESP). For ease of illustration, vehicle 120 will be used as an example in the following description, but it should be understood that the following description also applies to vehicles 130, 140, and 150, as well as any other vehicles.

[0026] For example, when vehicle 120 is driving in environments such as congested roads, icy roads, or rainy / snowy roads, it may frequently use the braking system and may encounter emergency situations, thereby triggering the vehicle stability system to control the braking system to prevent vehicle lock-up, skidding, etc. Therefore, the vehicle stability system of vehicle 120 needs to be accurately calibrated to ensure the safety of vehicle 120.

[0027] The embodiments of this disclosure can efficiently, automatically, cost-effectively, and accurately calibrate a vehicle's stability control system to ensure vehicle safety. The following will refer to... Figures 2 to 7 The example describes a method for calibrating a vehicle's body stability system according to this disclosure. Figure 2 The illustration shows a flowchart of a method 200 for calibrating a vehicle's body stability system according to an embodiment of the present disclosure. Figure 2 The method can be derived from Figure 1 The controller (e.g., ADAS) in the vehicle 120 (or any other vehicle) or any other controller, electronic device, or server that is in communication with the vehicle 120 (or any other vehicle) shall perform the operation.

[0028] like Figure 2 As shown, at box 202, vehicle model data is obtained. In some embodiments, vehicle model data may include physical parameters of the vehicle body, such as the vehicle's length, width, height, weight, and dimensions of related components (e.g., tires). In some embodiments, vehicle model data may also include vehicle type data. In some embodiments, type data may include the type of vehicle. In other embodiments, type data may include the clustering type under the vehicle type (which will be further described in the examples below).

[0029] At box 204, braking system data of the vehicle's braking system is acquired. In some embodiments, braking system data may be performance parameters (representing the braking capacity of the braking system) of relevant components of the vehicle's braking system (e.g., tires, wheel cylinders, body, and brake pedal), such as wheel speed range, wheel acceleration range, wheel cylinder pressure range, vehicle pitch angle range, vehicle steering angle range, and vehicle braking distance range. In some embodiments, two or more data items in the braking system data may be correlated, for example, wheel speed range and wheel cylinder pressure range may be correlated, but are not limited thereto. At box 206, based on vehicle model data and braking system data, a generative neural network model is used to generate parameters for calibrating the vehicle stability system.

[0030] In some embodiments, the generative neural network model can be a trained generative neural network model (the training process will be further described in the examples below). The trained generative neural network model can efficiently and automatically generate parameters for calibrating the vehicle stability system based solely on the vehicle's hardware parameters (the vehicle model data mentioned above) and related software parameters (the braking system data mentioned above), thereby improving the accuracy and efficiency of the generated parameters and saving significant costs in terms of manpower, time, vehicle wear and tear, and space.

[0031] The method according to embodiments of this disclosure can efficiently and automatically generate parameters for calibrating a vehicle's stability control system by using a generative neural network model based on vehicle-related data. This improves the accuracy and efficiency of vehicle stability control system calibration and reduces the cost of calibration. References below... Figures 3 to 7 This describes an example of using a generative neural network model to generate parameters for calibrating a vehicle stability system.

[0032] Figure 3 The illustration shows a flowchart of a process 300 for generating parameters according to an embodiment of the present disclosure. Figure 3 Process 300 corresponds to Figure 2 Box 206 in the middle. (For example...) Figure 3As shown, at box 302, it can be determined according to... Figure 2 The vehicle model data obtained at box 202 and in Figure 2 The braking system data acquired at box 204 is used to generate multiple sets of parameters using a generative neural network model. In some embodiments, each set of parameters may include multiple parameters, which may be parameters of multiple categories for calibrating the vehicle stability system, such as parameters related to anti-skid, parameters related to anti-lock braking, etc., so that each set of parameters forms a parameter combination, and the multiple sets of parameters provide a variety of options for parameter combinations.

[0033] At box 304, at least one set of parameters that satisfies the parameter constraints can be selected from multiple sets of parameters. In some embodiments, the parameter constraints may include individual constraints for at least one parameter in each set of parameters, and / or, the parameter constraints may include mutual constraints for at least two parameters in each set of parameters. In some embodiments, individual constraints may be used to limit the range of values ​​for individual parameters, etc. In some embodiments, mutual constraints may be used to limit the mutual constraints between parameters, for example, to limit the range of values ​​for the sum, difference, or ratio of two or more parameters. Thus, sets of parameters that do not meet the calibration requirements can be filtered out, making the calibration of the vehicle stability system more accurate, thereby ensuring vehicle safety.

[0034] At box 306, a set of parameters can be selected from at least one set of parameters based on predetermined filtering criteria to serve as parameters for calibrating the vehicle stability system. In some embodiments, the predetermined filtering criteria may be related to similarity (specifically...). Figure 4 The conditions described in the embodiments. In other embodiments, the predetermined filtering conditions may be user selection input by the user, such as the manufacturer or purchaser (e.g., its engineer) of the product corresponding to the method or apparatus of this application. The following will refer to... Figures 4 to 6 Let's further describe an example of selecting the above set of parameters.

[0035] It should be understood that if only one set of parameters is selected at box 304, then only that set of parameters can be selected at box 306 as the parameters used for calibrating the vehicle stability system, without judging the aforementioned predetermined screening conditions; or, box 306 can only provide the user with that set of parameters as candidate options, so that the user can only select that set of parameters as the parameters used for calibrating the vehicle stability system, without needing to filter according to the aforementioned predetermined screening conditions. Furthermore, if two or more sets of parameters are selected at box 304, then at box 306, the following can be referred to... Figures 4 to 6 The example shown illustrates how to select a set of parameters for calibrating the vehicle stability system.

[0036] Figure 4 The illustration shows a flowchart of a parameter selection process 400 according to an embodiment of the present disclosure. Figure 4 Process 400 can correspond to Figure 3 An example of box 306. For example... Figure 4 As shown, at box 402, it can be determined according to... Figure 3 At least one set of parameters selected at box 304 is used to obtain at least one set of theoretical driving data curves for the vehicle (e.g., theoretical curves of driving data such as speed, distance, and acceleration (but not limited to these) with respect to time). In some embodiments, the theoretical driving data curves may represent the expected driving data curves of the vehicle stability system for a corresponding set of parameters. For example, for each set of parameters, a set of theoretical driving data curves under different simulation conditions can be obtained by simulating the vehicle's driving in different environments using a first simulator. It should be understood that any first simulator capable of performing the above functions can be used to generate the theoretical driving data curves.

[0037] At box 404, respectively use in Figure 3 At box 304, each of the at least one set of parameters is selected to calibrate the vehicle stability system, thereby obtaining at least one set of actual driving data curves (e.g., actual curves of driving parameters such as speed, distance, and acceleration (but not limited to these) as a function of time) for the actual driving process of the vehicle. In other words, at box 404, each of the at least one set of parameters mentioned above can be used to perform the real-vehicle test described above, thereby obtaining the actual driving data curves corresponding to each set of parameters. It should be noted that the order of steps 402 and 404 is not limited here, and the two steps can be performed simultaneously.

[0038] At box 406, the similarity between each set of theoretical driving data curves obtained at box 402 and the corresponding set of actual driving data curves obtained at box 404 can be determined. In some embodiments, this similarity can indicate the similarity of shape between the theoretical driving data curves and the actual driving data curves. Therefore, in determining the similarity, minor differences in shape that may exist between the theoretical driving data curves and the actual driving data curves, such as translation, can be disregarded. In some embodiments, a second simulator (e.g., Calimaster) different from the first simulator described above can be used to determine the similarity, and the similarity can be quantified using a similarity score.

[0039] At box 408, it is possible to... Figure 3At box 304, the set of parameters corresponding to the highest similarity is selected from at least one set of parameters to serve as the parameters for calibrating the vehicle stability system. In some embodiments, the set of parameters corresponding to the highest similarity score determined by the second simulator in the above example can be selected as the parameters. Here, a specific set of parameters corresponds to a specific driving data curve (i.e., there is a correspondence between the two), and the driving data curve can represent the performance of the vehicle's braking system. Therefore, by detecting the similarity in shape between the obtained theoretical driving data curve and the actual driving curve, it can be determined whether the generated corresponding set of parameters meets the requirements (or the direction in which the parameters in the set need to be adjusted can be determined). This can greatly reduce the workload of real vehicle testing and improve the efficiency of generating parameters for calibrating the vehicle stability system.

[0040] Figure 5 The illustration shows an embodiment of the present disclosure. Figure 4 The diagram below illustrates the process 500, which uses a generative neural network model to generate parameters, corresponding to process 400. For example... Figure 5 As shown, in some embodiments, the generative neural network model 510 may include a generator model 511 and a discriminator model 512. The generative neural network model 510 may be a trained generative neural network model, which may be trained such that: the generator model 511 can generate corresponding parameters based on vehicle model data and braking system data as input (e.g., by constructing a vehicle dynamics model based on vehicle model data and braking system data to generate corresponding parameters), and the discriminator model 512 can verify the authenticity of the generated parameters, so that the generative neural network model 510 outputs parameters with authenticity greater than a predetermined authenticity threshold.

[0041] In some embodiments, the generative neural network model 510 may be a generative adversarial network (GAN). In other embodiments, the generative neural network model 510 may be a conditional tabular generative adversarial network (CTGAN). In still other embodiments, the generative neural network model 510 may be a diffusion model. It should be understood that the above are merely examples of the generative neural network model 510, and any other generative neural network model can be selected according to actual needs.

[0042] like Figure 5As shown, the generative neural network model 510 can receive input data 520, such as vehicle model data and braking system data. The generative neural network model 510 can generate multiple sets of parameters based on the input data 520. Figure 3 (Box 302). These multiple sets of parameters can be input to the first selection module 530 to select at least one set of parameters that satisfies the parameter constraints from the multiple sets of parameters. Figure 3 (Box 304). At least one set of parameters selected by the first selection module 530 can be input to the driving curve acquisition module 540 (e.g., the first simulator in the example above) to obtain at least one set of theoretical driving data curves for the vehicle. Figure 4 (frame 402).

[0043] Furthermore, at least one set of parameters selected by the first selection module 530 can also be calibrated to the vehicle stability system of the vehicle 120 to obtain at least one set of actual driving data curves for the actual driving process of the vehicle 120. Figure 4 (Box 404). The first module 551 in the second selection unit 550 (e.g., the second simulator in the example above) can determine the similarity between each set of theoretical driving data curves in the at least one set of theoretical driving data curves and the corresponding set of actual driving data curves in the at least one set of actual driving data curves. Figure 4 (Box 406). The second module 552 in the second selection unit 550 can select a set of parameters corresponding to the highest similarity from at least one set of parameters selected by the first selection module 530, as parameters for calibrating the vehicle stability system. Figure 4 (Box 408).

[0044] Through the above Figure 4 and Figure 5 The process shown allows for the selection of a set of parameters that most closely approximates the desired parameters of the vehicle stability system through a relatively small number of real-vehicle tests. This further improves the accuracy of vehicle stability system calibration and the efficiency of parameter determination, thereby enhancing vehicle safety.

[0045] The following is for reference Figure 6 To describe in Figure 3 In box 306, select one set of parameters from at least one selected set of parameters as another example of the parameter. Figure 6 A flowchart illustrating a process 600 for selecting parameters according to another embodiment of the present disclosure is shown. Figure 6 Process 600 can correspond to Figure 3 Box 306. (For example...) Figure 6 As shown, at box 602, a receiver can be received for... Figure 3The user's first input is selected at box 304, which selects at least one set of parameters. In some embodiments, the at least one set of parameters may be provided to the user by means of vision, hearing, etc., and the user's first input may be received through any input interface (e.g., keyboard, touch screen, or microphone, etc.).

[0046] At box 604, a set of parameters can be selected from at least one set of parameters based on the user's first input to serve as parameters for ESP calibration. This allows for further calibration of the vehicle's stability system based on the user's (e.g., the manufacturer or purchaser of the product corresponding to the method or device of this application, e.g., their engineer) preferences or experience. In other words, in this embodiment, parameters with improved accuracy and safety, and more suited to user preferences, can be generated as described above. It should be understood that... Figure 6 The process can also be used Figure 5 The process is implemented using the architecture shown in 500, the difference being... Figure 5 The driving curve acquisition unit 540 and the second selection unit 550 can be replaced with a user interaction module for interacting with the user to make a selection based on the user's first input.

[0047] Furthermore, in accordance with the above passage Figure 4 Process 400 or Figure 6 After selecting a set of parameters in process 600, these parameters can be further customized to better suit the user's preferences. Therefore, in some embodiments, method 200 according to this disclosure may further include: receiving a parameter for... Figure 4 In box 408 or in Figure 6 The second input is provided by a user (e.g., the aforementioned manufacturer or purchaser (e.g., its engineer)) who modifies one or more parameters from a set of parameters selected at box 604. For example, the set of parameters can be provided to the user visually, audibly, or otherwise. The second input can be received using an input interface that is the same as or different from the input interface used to receive the user's first input. The user is, for example, the manufacturer or purchaser (e.g., its engineer) of the product corresponding to the method or apparatus of this application.

[0048] In some embodiments, the method according to this disclosure may further include: modifying one or more parameters based on a second user input to obtain a modified set of parameters. In some embodiments, the method according to this disclosure may further include: using the modified set of parameters as parameters for calibrating the vehicle stability system. This allows for further customization of the vehicle stability system, making the calibrated system more compatible with the user's usage habits, thereby further improving the user's vehicle safety.

[0049] Furthermore, in some embodiments, in order to obtain the above references Figures 2 to 6 The described generative neural network model, according to the method 200 of the embodiments of this disclosure, may further include a process of training the generative neural network model. Figure 7 A flowchart illustrating a process 700 for training a generative neural network model according to an embodiment of the present disclosure is shown. Figure 7 As shown, at box 702, based on the vehicle model data of vehicle 120, multiple historical driving data curves of other vehicles with the same characteristics as vehicle 120, as well as braking system data of other vehicles, can be obtained (e.g., from a historical driving database for existing vehicle models). This vehicle model data can be compared with... Figure 2 The vehicle model data obtained at box 202 is the same as the vehicle model data, that is, the vehicle model data may include the physical parameters of the vehicle body of vehicle 120 and the type data of vehicle 120.

[0050] In some embodiments, type data may include the type of vehicle 120. In other embodiments, type data may include cluster types under the type of vehicle 120. In this case, in some embodiments, the aforementioned same features may be of the same type. In other embodiments, the aforementioned same features may be of the same cluster type. In some embodiments, when the type data includes cluster types under the type of vehicle 120 and the aforementioned same features are of the same cluster type, method 200 may further include: clustering vehicle model data and braking system data of multiple vehicles under each type to obtain at least one cluster type for each type. For example, for vehicle model data and braking system data under each type, vehicle model data and corresponding braking system data may be clustered into a predetermined number of cluster types according to the degree of difference between each vehicle model data and the corresponding braking system data. For example, in each cluster type, the difference between each vehicle model data is within a first predetermined difference range, and the difference between each braking system data is within a second predetermined difference range. For example, the number of cluster types obtained may be set according to actual needs; for example, several, a dozen, or dozens of cluster types may be obtained.

[0051] In some embodiments, the plurality of historical driving data curves can be actual driving data curves of the other vehicle under different historical parameters of the other vehicle's vehicle stability system and under different operating conditions. In some embodiments, each of the plurality of historical driving data curves can have a corresponding score, which can indicate the difference between each historical driving data curve and the desired driving data curve.

[0052] At box 704, a predetermined number of historical driving data curves that meet predetermined conditions can be selected from a plurality of historical driving data curves as annotation data. In some embodiments, a predetermined number of historical driving data curves with scores greater than a predetermined score threshold can be selected from a plurality of historical driving data curves as annotation data. It should be understood that the predetermined score threshold and the predetermined number can be determined according to actual needs.

[0053] At box 706, vehicle model data, braking system data, and labeled data from other vehicles can be used to train the generative neural network model. For example, during training, vehicle model data and braking system data from other vehicles can be input into the generative neural network model, and the output data can be parameters of the ESP system. These output parameters can be used to obtain corresponding driving data curves (e.g., theoretical or actual driving data curves for other vehicles obtained in a similar manner as described above). These driving data curves can be compared with corresponding labeled data to adjust the model parameters (e.g., the correlation functions of its generator and discriminator models) based on the comparison results (e.g., the differences between the corresponding driving data curves and the labeled data). Here, when determining the differences between the driving data curves and the corresponding labeled data, only the shape differences between the curves can be considered, without considering differences with minor shape effects such as translation.

[0054] In some embodiments, the above training can enable the generator model of the generative neural network model to generate corresponding parameters based on the input vehicle model data and braking system data, and the discriminator model of the generative neural network model to determine the authenticity of the generated parameters, so that the generative neural network model outputs parameters with authenticity greater than a predetermined authenticity threshold. It should be understood that this predetermined authenticity threshold can vary depending on actual needs and the specific generative neural network model used. In some embodiments, the above training can also ensure that the difference between the driving data curve corresponding to the parameters output by the generative neural network model and the corresponding labeled data is within a predetermined difference range. It should be understood that this predetermined difference range can be defined differently depending on actual needs.

[0055] Since a large amount of historical driving data curves have been accumulated for various types of vehicles, a large amount of comprehensive training data under various working conditions can be easily obtained during the training process. This enables efficient training of generative neural network models and allows the trained generative neural network models to have high accuracy.

[0056] Figure 8The illustration shows a schematic block diagram of an apparatus 800 for calibrating a vehicle's body stability system according to an embodiment of the present disclosure. Figure 8 As shown, the device 800 includes a first acquisition unit 810 configured to acquire vehicle model data. The device 800 also includes a second acquisition unit 820 configured to acquire braking system data of the vehicle's braking system. The device 800 further includes a parameter generation unit 830 configured to generate parameters for calibrating the vehicle stability system using a generative neural network model based on the vehicle model data and the braking system data.

[0057] In some embodiments, the parameter generation unit 830 may further be configured to generate multiple sets of parameters using a generative neural network model based on vehicle model data and braking system data, each set of parameters including multiple parameters. In some embodiments, the parameter generation unit 830 may further be configured to select at least one set of parameters from the multiple sets of parameters that satisfy parameter constraints. In some embodiments, parameter constraints may include individual constraints for at least one of the multiple parameters and / or mutual constraints for at least two of the multiple parameters. In some embodiments, the parameter generation unit 830 may further be configured to select a set of parameters from the at least one set of parameters as parameters according to predetermined filtering conditions.

[0058] In some embodiments, the parameter generation unit 830 may further be configured to obtain at least one set of theoretical driving data curves for the vehicle based on at least one set of parameters. In some embodiments, the parameter generation unit 830 may further be configured to calibrate the vehicle stability system using each set of parameters from the at least one set of parameters to obtain at least one set of actual driving data curves for the actual driving process of the vehicle. In some embodiments, the parameter generation unit 830 may further be configured to determine the similarity between each set of theoretical driving data curves in the at least one set of theoretical driving data curves and a corresponding set of actual driving data curves in the at least one set of actual driving data curves. In some embodiments, the parameter generation unit 830 may further be configured to select the set of parameters corresponding to the highest similarity from the at least one set of parameters as the parameters.

[0059] In some embodiments, the parameter generation unit 830 may also be configured to receive a first input from a user for selecting at least one set of parameters. In some embodiments, the parameter generation unit 830 may also be configured to select a set of parameters from at least one set of parameters as parameters based on the user's first input.

[0060] In some embodiments, the apparatus 800 may further include a modification unit configured to receive a second input from a user for modifying one or more parameters in a selected set of parameters. In some embodiments, the modification unit may also be configured to receive the second input from a user for modifying one or more parameters in a selected set of parameters. In some embodiments, the modification unit may also be configured to modify one or more parameters based on the user's second input to obtain a modified set of parameters. In some embodiments, the modification unit may also be configured to use the modified set of parameters as parameters.

[0061] In some embodiments, the device 800 may further include a training unit configured to obtain multiple historical driving data curves of other vehicles with the same characteristics as the vehicle, as well as braking system data of other vehicles, based on vehicle model data. In some embodiments, vehicle model data may include vehicle body physical parameters. In some embodiments, vehicle model data may further include vehicle type data. In some embodiments, type data may include the type of vehicle. In some embodiments, type data may include clustering types under vehicle types. In some embodiments, the aforementioned identical features may be of the same type. In other embodiments, the aforementioned identical features may be of the same clustering type. In some embodiments, when the type data includes clustering types under vehicle types and the identical features are of the same clustering type, the device 800 may further include a clustering unit configured to cluster vehicle model data and braking system data of multiple vehicles under each type to obtain at least one clustering type for each type.

[0062] In some embodiments, the multiple historical driving data curves can be actual driving data curves of other vehicles under different historical parameters of their vehicle stability systems and under different operating conditions (e.g., driving data collected by speed sensors, distance sensors, acceleration sensors, etc., and obtained through a time series model). In some embodiments, each of the multiple historical driving data curves may have a corresponding score. In some embodiments, the score may indicate the difference between each historical driving data curve and the desired driving data curve.

[0063] In some embodiments, the training unit may also be configured to select a predetermined number of historical driving data curves that meet predetermined conditions from a plurality of historical driving data curves as labeled data. In some embodiments, the training unit may also be configured to select a predetermined number of historical driving data curves with scores greater than a predetermined score threshold from a plurality of historical driving data curves as labeled data. In some embodiments, the training unit may also be configured to train a generative neural network model using vehicle model data of other vehicles, braking system data of other vehicles, and labeled data.

[0064] In some embodiments, the training of the training unit can enable: the generator model to generate corresponding parameters based on the input vehicle model data and braking system data; and the discriminator model to verify the authenticity of the generated parameters, so that the generative neural network model outputs parameters with authenticity greater than a predetermined authenticity threshold. In some embodiments, the training of the training unit can also enable: the difference between the driving data curve corresponding to the parameters output by the generative neural network model and the corresponding labeled data is within a predetermined difference range.

[0065] The apparatus 800 according to embodiments of this disclosure can efficiently generate parameters for calibrating a vehicle's stability control system (ESP) using a generative neural network model based on vehicle-related data. This improves the accuracy and efficiency of ESP calibration and reduces the cost of ESP calibration. The method and apparatus for calibrating a vehicle's stability control system provided by embodiments of this invention can generate calibration parameters for ESP systems under different vehicle models, scenarios, and operating conditions, thus expanding the range of calibration parameters. Parameters for calibrating a vehicle's stability control system can also be efficiently generated for new vehicles or extreme testing scenarios.

[0066] Figure 9 A schematic block diagram of an example device 900 suitable for implementing embodiments of the present disclosure is shown. The controller described above can be implemented using device 900. As shown, device 900 includes a processor 901 that can perform various appropriate actions and processes based on computer program instructions loaded into random access memory (RAM) 903 according to computer program instructions stored in read-only memory (ROM) 902. Various programs and data required for the operation of device 900 may also be stored in RAM 903. The processor 901, ROM 902, and RAM 903 are interconnected via bus 904. Input / output (I / O) interface 905 is also connected to bus 904.

[0067] The various processes and procedures described above, such as method 200 and processes 300 to 700, can be executed by processor 901. For example, in some embodiments, method 200 and processes 300 to 700 can be implemented as computer software programs tangibly contained in a machine-readable medium. In some embodiments, part or all of the computer program can be loaded and / or installed on device 900 via ROM 902. When the computer program is loaded into RAM 903 and executed by processor 901, one or more actions of method 200 and processes 300 to 700 described above can be performed.

[0068] This disclosure can be a method, apparatus, system, and / or computer program product. A computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for performing various aspects of this disclosure.

[0069] Computer-readable storage media can be tangible devices capable of holding and storing instructions for use by an instruction execution device. Computer-readable storage media can be, for example—but not limited to—electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or recessed protrusions storing instructions thereon, and any suitable combination of the foregoing. The computer-readable storage media used herein are not to be construed as transient signals themselves, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., light pulses through fiber optic cables), or electrical signals transmitted through wires.

[0070] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0071] Computer program instructions used to perform the operations of this disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing the status information of the computer-readable program instructions to implement various aspects of this disclosure.

Claims

1. A method (200) for calibrating a vehicle's body stability system, comprising: Obtain the vehicle model data of the vehicle mentioned in (202); Obtain braking system data of the braking system of the vehicle described in (204); as well as Based on the vehicle model data and the braking system data, a generative neural network model is used to generate (206) parameters for calibrating the vehicle stability system.

2. The method according to claim 1, wherein generating (206) parameters for calibrating the vehicle stability system using a generative neural network model based on the vehicle model data and the braking system data includes: Based on the vehicle model data and the braking system data, the generative neural network model is used to generate (302) multiple sets of parameters, each set of parameters including multiple parameters; Select (304) at least one set of parameters that satisfy the parameter constraints from the plurality of parameters, wherein the parameter constraints include: individual constraints for at least one of the plurality of parameters and / or mutual constraints for at least two of the plurality of parameters. According to predetermined screening conditions, select (306) a set of parameters from the at least one set of parameters as parameters for calibrating the vehicle stability system.

3. The method according to claim 2, wherein selecting (306) a set of parameters from the at least one set of parameters as the parameters according to predetermined screening conditions comprises: Based on the at least one set of parameters, obtain (402) at least one set of theoretical driving data curves for the vehicle; The vehicle stability system is calibrated using each of the at least one set of parameters to obtain (404) at least one set of actual driving data curves for the actual driving process of the vehicle. Determine the similarity between each set of theoretical driving data curves in the at least one set of theoretical driving data curves and the corresponding set of actual driving data curves in the at least one set of actual driving data curves; as well as Select (408) a set of parameters corresponding to the highest similarity from the at least one set of parameters as the parameters.

4. The method according to claim 2, wherein selecting (306) a set of parameters as the parameters from the at least one set of parameters according to predetermined screening conditions comprises: Receive first input from a user for selecting the at least one set of parameters (602); as well as Based on the user's first input, select (604) a set of parameters from the at least one set of parameters as the parameters.

5. The method according to claim 3 or 4, further comprising: Receive a second input from a user for modifying one or more parameters from the selected set of parameters; Based on the user's second input, the one or more parameters are modified to obtain the modified set of parameters; as well as The modified set of parameters is used as the parameters.

6. The method according to claim 1, further comprising: Based on the vehicle model data, obtain (702) multiple historical driving data curves of other vehicles with the same characteristics as the vehicle, as well as braking system data of the other vehicles; Select (704) a predetermined number of historical driving data curves that meet the predetermined conditions from the plurality of historical driving data curves as labeled data; as well as The generative neural network model is trained (706) using the vehicle model data of the other vehicles, the braking system data of the other vehicles, and the labeled data.

7. The method according to claim 6, wherein the plurality of historical driving data curves are actual driving data curves of the other vehicles under different historical parameters of the vehicle stability system of the other vehicles and under different operating conditions, and each of the plurality of historical driving data curves has a corresponding score, the score indicating the difference between each historical driving data curve and the desired driving data curve, and The selection of a predetermined number of historical driving data curves that meet predetermined conditions from the plurality of historical driving data curves as labeled data includes: The predetermined number of historical driving data curves with scores greater than a predetermined score threshold are selected from the plurality of historical driving data curves as the labeled data.

8. The method of claim 7, wherein the generative neural network model comprises a generator model and a discriminator model, and The training (706) enables the generator model to generate corresponding parameters based on the vehicle model data and braking system data as input, and the discriminator model to verify the authenticity of the generated parameters, so that the generative neural network model outputs parameters with authenticity greater than a predetermined authenticity threshold. The training also ensures that the difference between the driving data curve corresponding to the parameters output by the generative neural network model and the corresponding labeled data is within a predetermined range.

9. The method according to any one of claims 6 to 8, wherein the vehicle model data includes the vehicle's body physical parameters and the vehicle's type data. The type data includes at least one of the following: the type of the vehicle, or the clustering type under the type. The common features mentioned therein refer to the same type or the same cluster type, and Where the type data includes cluster types under the vehicle type, and the same features belong to the same cluster type, the method further includes: Clustering is performed on the vehicle model data and braking system data of multiple vehicles under each type to obtain at least one clustering type for each type.

10. An apparatus (800) for calibrating a vehicle stability system, comprising: The first acquisition unit (802) is configured to acquire vehicle model data; The second acquisition unit (804) is configured to acquire braking system data of the vehicle's braking system; The parameter generation unit (806) is configured to generate parameters for calibrating the vehicle stability system using a generative neural network model based on the vehicle model data and the braking system data.

11. A controller, comprising: At least one processor; as well as A memory coupled to the at least one processor and having instructions stored thereon, which, when executed by the at least one processor, cause the controller to perform the method according to any one of claims 1-9.

12. A vehicle comprising the controller according to claim 11.

13. A computer program product comprising computer-executable instructions, wherein the computer-executable instructions are executed by a processor to implement the method according to any one of claims 1 to 9.

Images