A method and device for calibrating brake ABS function based on neural network

Abstract

The present invention discloses a method and device for calibrating a braking ABS function based on a neural network. The method comprises: obtaining braking function data, vehicle model data and environmental data of a target vehicle; inputting the braking function data, vehicle model data and environmental data into a pre-trained ABS function calibration model to obtain calibration parameters and corresponding customer evaluations, wherein the ABS function calibration model is obtained by training an ANNs neural network with a training data set, and the calibration parameters include one of high-adhesion road surface parameters, low-adhesion road surface parameters, split road surface parameters, butt road surface parameters, turning parameters, and noisy road surface parameters, as well as motor speed control parameters. The present invention predicts calibration parameters and customer evaluations by means of deep learning, and compared with the manual calibration method, can shorten the calibration work cycle, improve the calibration work efficiency, and reduce the influence of subjective factors.

Description

Technical Field

[0001] The invention belongs to the technical field of vehicles, and in particular relates to a braking ABS function calibration method and device based on a neural network. Background Art

[0002] The main purpose of automobile brake system software calibration is to determine the control parameters of the brake system through actual vehicle road tests. On the one hand, the corresponding test method needs to use some additional equipment to obtain accurate measurement results. On the other hand, the consideration of parameter robustness in the orthogonality test design of brake system calibration needs to be improved.

[0003] The calibration work usually starts with determining the calibration test plan and testing the vehicle braking performance based on different road conditions. The wheel speed, vehicle speed and other information are collected through actual vehicle testing and compared with the brake system calibration test standard data. After the performance characterization meets the requirements, the comfort and satisfaction evaluation is carried out.

[0004] The correction work of the brake parameters in the existing scheme comes from the work of the calibrators. The complexity of the brake calibration parameters makes it difficult to accurately locate the required correction parameters through the data collection and analysis of the brake function. This part of the work is strongly related to the work experience of the tester. Therefore, for less experienced calibrators, the calibration of the brake system parameters often takes a lot of time. In addition, the comfort evaluation in the evaluation indicators of the existing scheme is a person's subjective feeling. There is a phenomenon that the evaluation standards of the calibrators and customers are inconsistent, resulting in low accuracy of the calibration results. Summary of the invention

[0005] An object of the present invention is to provide a method and device for calibrating a braking ABS function based on a neural network, which can solve the technical problems in the prior art that manual calibration takes a long time and the accuracy of the calibration result is low.

[0006] In a first aspect, the present invention provides a brake ABS function calibration method based on a neural network, comprising:

[0007] Obtaining the braking function data, vehicle model data and environmental data of the target vehicle, wherein the braking function data includes initial wheel speed, braking distance, deceleration, slip rate, offset rate, yaw rate, and pedal travel; the vehicle model data includes vehicle mass, wheelbase, front track, rear track, front axle load, rear axle load, number of teeth on the front gear ring, number of teeth on the rear gear ring, tire rolling radius, tire pressure, and brake type; and the environmental data includes weather and ambient temperature;

[0008] The braking function data, the vehicle model data and the environmental data are input into a pre-trained ABS function calibration model to obtain calibration parameters and corresponding customer evaluations, wherein the ABS function calibration model is obtained by training an ANNs neural network with a data set to be trained, and the calibration parameters include one of high-adhesion road surface parameters, low-adhesion road surface parameters, split road surface parameters, butt road surface parameters, turning parameters, noisy road surface parameters and motor speed control parameters.

[0009] Optionally, the ABS function calibration model includes a first neural network and a second neural network, the first neural network is used to predict the calibration personnel evaluation, pedal feedback and working noise, and transmit the prediction result to the second neural network;

[0010] The second neural network is a two-head model of conditional judgment, which predicts calibration parameters and customer evaluations by identifying calibration test items.

[0011] Optionally, the input of the first neural network includes the brake function data, the vehicle model data and the environmental data, and the output of the first neural network includes calibration personnel evaluation, pedal feedback and working noise;

[0012] The input of the second neural network includes the pedal feedback, the working noise and the brake function data, and the output of the second neural network is the calibration parameters and the customer evaluation.

[0013] Optionally, the first neural network includes 3 hidden layers, and each hidden layer of the first neural network includes 60 neurons;

[0014] The second neural network includes 5 hidden layers, and each hidden layer of the second neural network includes 40 neurons.

[0015] Optionally, the neuron activation function is a Sigmoid function, and the loss function is a mean square error.

[0016] Optionally, the method further comprises:

[0017] Set the evaluation factor set and evaluation set ;

[0018] ;

[0019] ;

[0020] in, For deceleration evaluation, For slip rate evaluation, For the deviation rate evaluation, For braking distance evaluation, For pedal feedback evaluation, For the evaluation of working noise, It means excellence, Indicates qualified, Indicates failure to meet the standards;

[0021] According to the evaluation factor set and evaluation set Establishing fuzzy evaluation matrix , the fuzzy evaluation matrix is ​​used to determine the calibration personnel evaluation and the customer evaluation.

[0022] Optionally, before inputting the braking function data and the environmental data into a pre-trained ABS function calibration model, the method further comprises:

[0023] Acquire a training data set, the training data set including calibration parameters of a plurality of vehicles and corresponding calibration personnel evaluations and customer evaluations;

[0024] The ANNs neural network is trained according to the training data set to obtain the ABS function calibration model.

[0025] In a second aspect, the present invention provides a device using the neural network-based braking ABS function calibration method described in the first aspect of the present invention, comprising:

[0026] an acquisition module, for acquiring the braking function data, vehicle model data and environmental data of the target vehicle, wherein the braking function data includes initial wheel speed, braking distance, deceleration, slip rate, offset rate, yaw rate and pedal travel; the vehicle model data includes vehicle mass, wheelbase, front track, rear track, front axle load, rear axle load, number of teeth on the front gear ring, number of teeth on the rear gear ring, tire rolling radius, tire pressure and brake type; and the environmental data includes weather and ambient temperature;

[0027] An input module is used to input the braking function data, the vehicle model data and the environmental data into a pre-trained ABS function calibration model to obtain calibration parameters and corresponding customer evaluations, wherein the ABS function calibration model is obtained by training an ANNs neural network with a data set to be trained, and the calibration parameters include one of high-adhesion road parameters, low-adhesion road parameters, split road parameters, butt road parameters, turning parameters, noisy road parameters and motor speed control parameters.

[0028] In a third aspect, the present invention provides an electronic device comprising a processor and a memory, wherein the memory stores programs or instructions that can be run on the processor, and when the program or instructions are executed by the processor, the steps of a neural network-based braking ABS function calibration method as described in the first aspect of the present invention are implemented.

[0029] In a fourth aspect, the present invention provides a readable storage medium storing a program or instruction. When the program or instruction is executed by a processor, the steps of a neural network-based braking ABS function calibration method as described in the first aspect of the present invention are implemented.

[0030] The beneficial effects of the present invention are as follows: the present invention predicts calibration parameters and customer evaluations by means of deep learning, which can shorten the calibration cycle, improve the calibration efficiency, and reduce the influence of subjective factors compared with the manual calibration method. In addition, the present invention uses a variety of vehicle model data, brake function data, and environmental data for calibration, and improves the accuracy of calibration parameters by using as many types of data as possible. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 It is a schematic diagram of the ABS function calibration process in the prior art.

[0032] Figure 2 It is a flow chart of a braking ABS function calibration method based on neural network in an embodiment of the present invention.

[0033] Figure 3 Schematic diagram of the ABS function calibration model in an embodiment of the present invention.

[0034] Figure 4 It is a schematic diagram of a braking ABS function calibration method based on a neural network in another embodiment of the present invention. DETAILED DESCRIPTION

[0035] Various exemplary embodiments of the present invention will now be described in detail with reference to the accompanying drawings. It should be noted that the relative arrangement of components and steps, numerical expressions and numerical values ​​set forth in these embodiments do not limit the scope of the present invention unless otherwise specifically stated.

[0036] The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the invention, its application, or uses.

[0037] The techniques, methods and devices known to those of ordinary skill in the relevant art may not be discussed in detail, but where appropriate, the techniques, methods and devices should be considered as part of the specification. In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limiting. Therefore, other examples of the exemplary embodiments may have different values.

[0038] It should be noted that like reference numerals and letters refer to similar items in the following figures, and therefore, once an item is defined in one figure, it need not be further discussed in subsequent figures.

[0039] In the specification and claims of the present invention, if the term "first" or "second" is involved, it may explicitly or implicitly include one or more of the features. In the description of the present invention, unless otherwise specified, "plurality" means two or more. In addition, "and / or" in the specification and claims means at least one of the connected objects, and the character " / " generally means that the related objects are in an "or" relationship.

[0040] The main purpose of automobile brake system software calibration is to determine the control parameters of the brake system through actual vehicle road tests. On the one hand, the corresponding test method needs to use some additional equipment to obtain accurate measurement results. On the other hand, the consideration of parameter robustness in the orthogonality test design of brake system calibration needs to be improved.

[0041] The calibration work usually starts with determining the calibration test plan and testing the vehicle braking performance based on different road conditions. Figure 1 The figure shows the conventional brake system calibration test process. For different test items (function, road information, etc.), wheel speed, vehicle speed and other information are collected through real vehicle testing and compared with the brake system calibration test standard data. After the performance characterization meets the requirements, the comfort and satisfaction evaluation is carried out.

[0042] In the existing solution, the correction of brake parameters comes from the work of calibrators. The complexity of brake calibration parameters makes it difficult to accurately locate the required correction parameters through data collection and analysis of brake functions. This part of the work is strongly related to the work experience of testers. Therefore, for less experienced calibrators, the calibration of brake system parameters often takes a lot of time.

[0043] In addition, during the calibration process, the calibration staff of the manufacturer usually evaluates the calibration parameters first. When the calibration parameters meet the requirements, they will be delivered to the customer for final evaluation. The comfort evaluation in the evaluation indicators of the existing scheme is a person's subjective feeling, and there is a phenomenon that the evaluation standards of the calibration staff and the customers are inconsistent.

[0044] like Figure 2 As shown, this embodiment introduces a braking ABS function calibration method based on a neural network, including steps 2100-2200.

[0045] Step 2100: Obtain the braking function data, vehicle model data and environmental data of the target vehicle, wherein the braking function data includes initial wheel speed, braking distance, deceleration, slip rate, offset rate, yaw angular velocity, and pedal travel; the vehicle model data includes vehicle mass, wheelbase, front track, rear track, front axle load, rear axle load, number of front gear teeth, number of rear gear teeth, tire rolling radius, tire pressure, and brake type; and the environmental data includes weather and ambient temperature.

[0046] The target vehicle is a vehicle that needs to be calibrated for the ABS (Antilock Brake System) function. During the calibration process, it is often necessary to calibrate various models of vehicles, and the model data of different models of vehicles are different, which will affect the final calibration parameters. For different models of vehicles, the calibration parameters are different even in the same scenario.

[0047] In addition, environmental data also affects the calibration parameters. Environmental data includes weather and ambient temperature. For example, sunny, rainy, snowy and other weather conditions. When the weather is sunny, the road conditions are good. When it rains or snows, the vehicle is prone to slipping.

[0048] The braking function data indicates the driving status of the vehicle, and different statuses correspond to different calibration parameters.

[0049] Step 2200: Input the braking function data, the vehicle model data and the environmental data into a pre-trained ABS function calibration model to obtain calibration parameters and corresponding customer evaluations, wherein the ABS function calibration model is obtained by training an ANNs (Artificial Neural Network) neural network with a training data set, and the calibration parameters include one of high-adhesion road parameters, low-adhesion road parameters, split road parameters, butt road parameters, turning parameters, noisy road parameters and motor speed control parameters.

[0050] The present invention establishes an ANNs neural network model based on Python & TensorFlow on the server, and performs sample set training, where the sample data set comes from real vehicle calibration. The trained model is tested with prediction accuracy as the evaluation index.

[0051] During the actual vehicle calibration process, the trained ABS function calibration model is deployed on the calibrator's computer to participate in the vehicle calibration task. The calibrator inputs various parameters to correct the ABS function calibration parameters.

[0052] Since vehicles may travel on various road surfaces, different parameters need to be calibrated for different road surfaces, such as high-adhesion road surfaces, low-adhesion road surfaces, split road surfaces, butt road surfaces, turning road surfaces, noisy road surfaces, etc. The ABS function calibration model can output corresponding parameters according to actual needs.

[0053] Customer evaluation is the customer's evaluation result of the calibration parameters predicted by the ABS functional calibration model, indicating the possibility of the calibration result passing the customer's acceptance. When the possibility is higher than a certain level, it means that the calibration result can be delivered to the customer for acceptance.

[0054] The present invention predicts calibration parameters and customer evaluations through deep learning. Compared with manual calibration, it can shorten the calibration cycle, improve calibration efficiency, and reduce the influence of subjective factors. In addition, the present invention uses multiple vehicle model data, brake function data, and environmental data for calibration, and improves the accuracy of calibration parameters through as many types of data as possible.

[0055] In this embodiment, the ABS function calibration model includes a first neural network and a second neural network. The first neural network is used to predict the evaluation of the calibration personnel, pedal feedback and working noise, and transmit the prediction results to the second neural network; the second neural network is a conditional judgment double-headed model, which predicts calibration parameters and customer evaluations by identifying calibration test items.

[0056] like Figure 3 As shown, the ABS function calibration model includes two ANN neural networks, namely the first neural network and the second neural network. The first neural network is used to predict the evaluation of the calibration personnel, pedal feedback and working noise. The second neural network is used to predict the calibration parameters and customer evaluation.

[0057] like Figure 4 As shown in the figure, the calibration test items represent the road scenes corresponding to the calibration parameters, such as high-adhesion road surface, low-adhesion road surface, split road surface, butt road surface, turn, noisy road surface, etc. The calibration test items are input by the calibration personnel during the calibration process. The calibration personnel can set the corresponding calibration test items according to actual needs, so that the ABS function calibration model can output the corresponding calibration parameters.

[0058] In this embodiment, the input of the first neural network includes the braking function data, the vehicle model data and the environmental data, and the output of the first neural network includes the calibration personnel evaluation, pedal feedback, and working noise; the input of the second neural network includes the pedal feedback, the working noise and the braking function data, and the output of the second neural network is the calibration parameters and the customer evaluation.

[0059] The ABS function of the braking system is calibrated as a highly nonlinear behavior. The data set of the two-layer neural network is divided into three parts, and the variable names are shown in Table 1. Table 1 includes input parameters and output parameters.

[0060] Table 1 List of input and output parameters of ABS function calibration model

[0061]

[0062] The input of the first neural network consists of the input parameters numbered 1-20 in Table 1. Figure 3 As shown, the input layer InV1 of the first neural network contains 20 neurons, corresponding to 20 input parameters. The variables numbered 1-11 are vehicle model data, the variables numbered 12 and 13 are environmental data, and the variables numbered 14-20 are brake function data. The vehicle model data and environmental data can be obtained before the experiment, and the brake function data can be obtained through various sensors of the vehicle during the experiment. The output parameters numbered 21, 22, and 23 in Table 1 are the outputs of the first neural network. The input of the first neural network is recorded as , expressed as:

[0063] .

[0064] in, They represent the input parameters numbered 1-20 in Table 1 respectively.

[0065] The output of the first neural network is recorded as , expressed as:

[0066] .

[0067] Referring to Table 1, IE is the evaluation of the calibration personnel, PF is the pedal feedback, and Noi is the working noise.

[0068] The pedal feedback and working noise output by the first neural network are used as the input of the second neural network. Meanwhile, the input of the second neural network also includes the input parameters numbered 14-20 in Table 1. Figure 3 As shown, the input layer InV2 of the second neural network includes 9 neurons, of which the first neuron and the second neuron correspond to the pedal feedback and working noise output by the first neural network respectively, and the remaining neurons correspond to the input parameters numbered 14-20. The input of the second neural network is recorded as , expressed as:

[0069] .

[0070] The output of the second neural network is recorded as , expressed as:

[0071] .

[0072] Among them, i is the value of the calibration test item t, and the value range is 1-6. The high-adhesion road parameter Y1, low-adhesion road parameter Y2, split road parameter Y3, and butt road parameter Y4 contain 32 elements, the turning parameter Y5 contains 17 elements, and the noisy road parameter Y6 contains 16 elements. The motor speed control parameter EM includes 16 elements.

[0073] like Figure 3 As shown, when the calibration test item t takes a value of 1, the output of the second neural network includes high-adhesion road parameter Y1, customer evaluation OE, and motor speed parameter EM. When the calibration test item t takes a value of 2, the output of the second neural network includes low-adhesion road parameter Y2, customer evaluation OE, and motor speed parameter EM. When the calibration test item t takes a value of 6, the output of the second neural network includes noisy road parameter Y6, customer evaluation OE, and motor speed parameter EM.

[0074] In this embodiment, the first neural network includes 3 hidden layers, and each hidden layer of the first neural network includes 60 neurons; the second neural network includes 5 hidden layers, and each hidden layer of the second neural network includes 40 neurons.

[0075] In this embodiment, the neuron activation function is the Sigmoid function, and the loss function is the mean square error. The error limit is 10 -3 .

[0076] The predicted value of the neural network output It can be expressed as:

[0077] .

[0078] in, is the weight matrix, is the bias vector. Update and optimize the weight matrix and bias vector of the neural network.

[0079] In this embodiment, the method further includes: setting an evaluation factor set and evaluation set According to the evaluation factor set and evaluation set Establishing fuzzy evaluation matrix , the fuzzy evaluation matrix is ​​used to determine the calibration personnel evaluation and the customer evaluation.

[0080] .

[0081] .

[0082] in, For deceleration evaluation, For slip rate evaluation, For the deviation rate evaluation, For braking distance evaluation, For pedal feedback evaluation, For the evaluation of working noise, It means excellence, Indicates qualified, Indicates unqualified.

[0083] The predicted output and It contains two subjective evaluation indicators IE and OE, which need to be quantified before training the neural network.

[0084] The evaluation factor set includes deceleration evaluation, slip rate evaluation, offset rate evaluation, braking distance evaluation, pedal feedback evaluation, and working noise evaluation. Among them, deceleration, slip rate, offset rate, and braking distance are observed quantities, which can be used for data collection and related numerical calculations through calibration tests, and scored according to specific standards. It can include industry general standards, such as GB21670-2008, GB12676-2014, ECE R13 and other standards. It can also include manufacturer-defined internal calibration standards. Pedal feedback evaluation and working noise evaluation come from the scores of calibration personnel or acceptance customers.

[0085] Three evaluation results are set in the evaluation set, namely excellent, qualified and unqualified. Specifically, the evaluation results of each indicator can be quantified, and corresponding judgment thresholds can be set for various evaluation results. For example, the evaluation results are 0-100. When the evaluation result of a certain indicator is 100, it is excellent. When the evaluation result of a certain indicator is 80-99, it is qualified. And when the evaluation result of a certain indicator is 0-79, it is unqualified.

[0086] A fuzzy evaluation matrix is ​​established according to the evaluation factor set and the evaluation set to determine the comprehensive evaluation model of the subjective evaluation indicators IE and OE. During the neural network training process, the evaluation indicators IE and OE are determined through the fuzzy evaluation matrix.

[0087] In this embodiment, before inputting the braking function data and the environmental data into a pre-trained ABS function calibration model, the method further includes: obtaining a training data set, the training data set including calibration parameters of multiple vehicles and corresponding calibration personnel evaluations and customer evaluations; training the ANNs neural network according to the training data set to obtain the ABS function calibration model.

[0088] The training data set is derived from manual calibration. After obtaining the training data set, an initial ANNs neural network model is created. The initial ANNs neural network model includes two layers of neural networks. The structure of each neural network refers to the relevant introduction in the previous embodiment and will not be repeated here. The initial ANNs neural network model is initialized, the weights and biases of the neural network are initialized, and then the training data set is used for training to obtain the ABS function calibration model.

[0089] This embodiment introduces a braking ABS function calibration device based on a neural network, including:

[0090] an acquisition module, for acquiring the braking function data, vehicle model data and environmental data of the target vehicle, wherein the braking function data includes initial wheel speed, braking distance, deceleration, slip rate, offset rate, yaw rate and pedal travel; the vehicle model data includes vehicle mass, wheelbase, front track, rear track, front axle load, rear axle load, number of teeth on the front gear ring, number of teeth on the rear gear ring, tire rolling radius, tire pressure and brake type; and the environmental data includes weather and ambient temperature;

[0091] An input module is used to input the braking function data, the vehicle model data and the environmental data into a pre-trained ABS function calibration model to obtain calibration parameters and corresponding customer evaluations, wherein the ABS function calibration model is obtained by training an ANNs neural network with a data set to be trained, and the calibration parameters include one of high-adhesion road parameters, low-adhesion road parameters, split road parameters, butt road parameters, turning parameters, noisy road parameters and motor speed control parameters.

[0092] This embodiment introduces an electronic device, including a processor and a memory, wherein the memory stores programs or instructions that can be run on the processor, and when the program or instructions are executed by the processor, the steps of a neural network-based braking ABS function calibration method as described in any embodiment of the present invention are implemented.

[0093] This embodiment introduces a readable storage medium, on which a program or instruction is stored. When the program or instruction is executed by a processor, the steps of a neural network-based braking ABS function calibration method as described in any embodiment of the present invention are implemented.

[0094] Although some specific embodiments of the present invention have been described in detail by way of example, it will be appreciated by those skilled in the art that the above examples are for illustration only and are not intended to limit the scope of the present invention. It will be appreciated by those skilled in the art that the above embodiments may be modified without departing from the scope and spirit of the present invention. The scope of the present invention is defined by the appended claims.

[0095] Those of ordinary skill in the art will appreciate that the modules and algorithm steps described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0096] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described devices and equipment can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0097] In the embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the modules is only a logical function division. There may be other division methods in actual implementation, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or modules, which can be electrical, mechanical or other forms.

[0098] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed on multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the embodiments of the present invention.

[0099] In addition, each functional module in the embodiment of the present invention may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0100] If the functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods of various embodiments of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, ROM, RAM, magnetic disks, or optical disks.

[0101] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the inventive concept. For example, the above features are replaced with (but not limited to) technical features with similar functions disclosed in the present application.

[0102] It should be understood that the size of the sequence number of each step in the content of the invention and the embodiments of the present invention does not absolutely mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention. For the purpose of example and description, the foregoing description of the implementation of the present disclosure has been given. The foregoing description is not exhaustive and is not intended to limit the present disclosure to the exact form disclosed. Various deformations and modifications may exist according to the above teachings, or various deformations and modifications may be obtained from the practice of the present disclosure. These embodiments are selected and described to illustrate the principles of the present disclosure and its practical application, so that those skilled in the art can use the present disclosure in various embodiments and various modifications suitable for the specific purpose conceived.

Claims

1. A method for calibrating braking ABS function based on neural network, characterized in that: include: Obtaining the braking function data, vehicle model data and environmental data of the target vehicle, wherein the braking function data includes initial wheel speed, braking distance, deceleration, slip rate, offset rate, yaw rate, and pedal travel; the vehicle model data includes vehicle mass, wheelbase, front track, rear track, front axle load, rear axle load, number of teeth on the front gear ring, number of teeth on the rear gear ring, tire rolling radius, tire pressure, and brake type; and the environmental data includes weather and ambient temperature; Inputting the braking function data, the vehicle model data and the environmental data into a pre-trained ABS function calibration model to obtain calibration parameters and corresponding customer evaluations, wherein the ABS function calibration model is obtained by training an ANNs neural network with a training data set, and the calibration parameters include one of a high-adhesion road surface parameter, a low-adhesion road surface parameter, a split road surface parameter, a butt road surface parameter, a turning parameter, a noisy road surface parameter and a motor speed control parameter; The ABS function calibration model includes a first neural network and a second neural network, wherein the first neural network is used to predict the calibration personnel evaluation, pedal feedback and working noise, and transmit the prediction result to the second neural network; The second neural network is a two-head model of conditional judgment, which predicts calibration parameters and customer evaluations by identifying calibration test items.

2. The method according to claim 1, characterized in that The input of the first neural network includes the brake function data, the vehicle model data and the environmental data, and the output of the first neural network includes calibration personnel evaluation, pedal feedback and working noise; The input of the second neural network includes the pedal feedback, the working noise and the brake function data, and the output of the second neural network is the calibration parameters and the customer evaluation.

3. The method according to claim 2, characterized in that The first neural network includes 3 hidden layers, and each hidden layer of the first neural network includes 60 neurons; The second neural network includes 5 hidden layers, and each hidden layer of the second neural network includes 40 neurons.

4. The method according to claim 3, characterized in that The neuron activation function is the Sigmoid function and the loss function is the mean square error.

5. The method according to claim 1, characterized in that The method further comprises: Set the evaluation factor set and evaluation set ; ； ； in, For deceleration evaluation, For slip rate evaluation, For the deviation rate evaluation, For braking distance evaluation, For pedal feedback evaluation, For the evaluation of working noise, It means excellence, Indicates qualified, Indicates failure to meet the standards; According to the evaluation factor set and evaluation set Establishing fuzzy evaluation matrix , the fuzzy evaluation matrix is ​​used to determine the calibration personnel evaluation and the customer evaluation.

6. The method according to claim 1, characterized in that Before inputting the braking function data and the environmental data into a pre-trained ABS function calibration model, the method further includes: Acquire a training data set, the training data set including calibration parameters of a plurality of vehicles and corresponding calibration personnel evaluations and customer evaluations; The ANNs neural network is trained according to the training data set to obtain the ABS function calibration model.

7. A device using a neural network-based braking ABS function calibration method as claimed in any one of claims 1 to 6, characterized in that: include: an acquisition module, for acquiring the braking function data, vehicle model data and environmental data of the target vehicle, wherein the braking function data includes initial wheel speed, braking distance, deceleration, slip rate, offset rate, yaw rate and pedal travel; the vehicle model data includes vehicle mass, wheelbase, front track, rear track, front axle load, rear axle load, number of teeth on the front gear ring, number of teeth on the rear gear ring, tire rolling radius, tire pressure and brake type; and the environmental data includes weather and ambient temperature; An input module is used to input the braking function data, the vehicle model data and the environmental data into a pre-trained ABS function calibration model to obtain calibration parameters and corresponding customer evaluations, wherein the ABS function calibration model is obtained by training an ANNs neural network with a data set to be trained, and the calibration parameters include one of high-adhesion road parameters, low-adhesion road parameters, split road parameters, butt road parameters, turning parameters, noisy road parameters and motor speed control parameters.

8. An electronic device, characterized in that: It includes a processor and a memory, wherein the memory stores a program or instruction that can be run on the processor, and when the program or instruction is executed by the processor, the steps of a neural network-based braking ABS function calibration method as described in any one of claims 1 to 6 are implemented.

9. A readable storage medium, characterized in that: The readable storage medium stores a program or instruction, and when the program or instruction is executed by the processor, the steps of a neural network-based braking ABS function calibration method as described in any one of claims 1 to 6 are implemented.

Images