Braking control method and system based on deep learning, storage medium and equipment

Abstract

The invention discloses a braking control method and system based on deep learning, a storage medium and equipment, and the method comprises the steps: collecting first vehicle state data through a preset sensor in the historical driving process of a vehicle; performing data processing on the first vehicle state data, and importing the first vehicle state data into a pre-provided deep learning model for model training to obtain a target model; in the running process of the vehicle, second vehicle state data are collected in real time through a preset sensor; and importing the second vehicle state data into the target model, and outputting a brake control strategy to control a brake actuator of the vehicle to work according to the brake control strategy. According to the method, the first vehicle state data is collected in the vehicle driving process to train the deep learning model, then the second vehicle state data is imported into the target model, the brake control strategy can be output, accurate control over the brake process is achieved, and the comfort degree and the satisfaction degree of a user are improved.

Description

Technical Field

[0001] The present invention relates to the field of vehicle control technology, and in particular to a brake control method, system, storage medium and device based on deep learning. Background Art

[0002] In recent years, with the rapid development of artificial intelligence technology, deep learning algorithms have achieved remarkable application results in many fields.

[0003] In the field of vehicle control, although some control systems based on deep learning have been proposed, most of these systems focus on vehicle driving stability, energy saving and other aspects, and there is relatively little research on the control of the brake system. At present, the relevant implementation schemes are mainly focused on brake control based on sensor data, but these methods often ignore the comfort needs of passengers, and their adaptability to assisted driving needs to be improved.

[0004] Therefore, the prior art still has the technical problem that the braking force cannot be accurately controlled during braking, resulting in a poor passenger experience. Summary of the invention

[0005] In view of the shortcomings of the prior art, the purpose of the present invention is to provide a brake control method, system, storage medium and device based on deep learning, aiming to solve the technical problem that the prior art cannot achieve accurate control of the braking force during braking, resulting in poor passenger experience.

[0006] A first aspect of the present invention is to provide a brake control method based on deep learning, the method comprising:

[0007] During the historical driving process of the vehicle, first vehicle status data is collected through a preset sensor, wherein the first vehicle status data includes driving speed, braking force and acceleration;

[0008] Performing data processing on the first vehicle state data, importing the first vehicle state data into a pre-provided deep learning model for model training, and obtaining a target model;

[0009] During the driving process of the vehicle, the second vehicle status data is collected in real time through a preset sensor, and the second vehicle status data includes driving speed, braking force and acceleration;

[0010] The second vehicle state data is imported into the target model, and a brake control strategy is output to control the brake actuator of the vehicle to work according to the brake control strategy.

[0011] According to one aspect of the above technical solution, the step of processing the first vehicle status data, importing the first vehicle status data into a pre-provided deep learning model for model training, and obtaining a target model includes:

[0012] Loading, cleaning and converting the first vehicle status data, and converting the first vehicle status data into a data format that meets preset requirements;

[0013] The preprocessed first vehicle state data is used as input data and loaded into the deep learning model in batches through a data loader;

[0014] Forward propagating the first vehicle state data through the deep learning model to calculate model output data of the deep learning model;

[0015] Using a cross entropy loss function to calculate the error between the model output data and the target value;

[0016] According to the gradient of the cross entropy loss function, back propagation is performed to calculate the gradient of each parameter in the deep learning model;

[0017] Use an optimizer to update the corresponding parameters of the deep learning model according to the gradient of each parameter to obtain a target model.

[0018] According to one aspect of the above technical solution, during the driving process of the vehicle, a second vehicle status data is collected in real time through a preset sensor, and the second vehicle status data includes a driving speed, a braking force and an acceleration step, including:

[0019] During the driving process of the vehicle, the vehicle's driving speed, braking force and acceleration are collected in real time through preset sensors to obtain second vehicle status data;

[0020] Preprocessing the second vehicle status data, including filtering and denoising, and performing format conversion to obtain the second vehicle status data in JSON format;

[0021] The second vehicle status data is transmitted to a central processor or a data storage unit via a CAN bus of the vehicle.

[0022] According to one aspect of the above technical solution, the step of importing the second vehicle state data into the target model and outputting a brake control strategy to control the brake actuator of the vehicle to work according to the brake control strategy includes:

[0023] calling the target model and continuously importing the continuously generated second vehicle state data into the target model;

[0024] Controlling the target model to output data parameters of braking timing node, braking force and braking duration according to the second vehicle state data to obtain a braking control strategy;

[0025] Performing data conversion on the brake control strategy to obtain a control signal for controlling a brake actuator;

[0026] The control signal is transmitted to the brake actuator via the CAN bus, so that the brake actuator performs corresponding braking operations according to the instructions of the control signal.

[0027] According to one aspect of the above technical solution, the method further includes:

[0028] When the brake actuator performs a corresponding brake operation according to the instruction of the control signal, the state of the brake actuator is monitored in real time by a preset brake sensor to obtain monitoring data;

[0029] The brake control strategy is adjusted in real time according to the monitoring data.

[0030] According to one aspect of the above technical solution, the step of adjusting the brake control strategy in real time according to the monitoring data includes:

[0031] Comparing the monitoring data with the expected effect data of the brake control strategy to determine whether there is a deviation or anomaly;

[0032] If so, evaluating the expected effect data of the current brake control strategy, identifying the parameters that need to be adjusted in the brake control strategy based on the evaluation results, and obtaining an adjustment strategy;

[0033] The brake control strategy is adjusted in real time according to the adjustment strategy.

[0034] A second aspect of the present invention is to provide a brake control system based on deep learning, which is applied to the method described in the above technical solution, and the system comprises:

[0035] A first data acquisition module, used to collect first vehicle status data through a preset sensor during the historical driving process of the vehicle, wherein the first vehicle status data includes driving speed, braking force and acceleration;

[0036] A model training module, used to process the first vehicle state data, import the first vehicle state data into a pre-provided deep learning model for model training, and obtain a target model;

[0037] A second data acquisition module, used to collect second vehicle status data in real time through a preset sensor during the driving process of the vehicle, wherein the second vehicle status data includes driving speed, braking force and acceleration;

[0038] The strategy output module is used to import the second vehicle state data into the target model and output a brake control strategy to control the vehicle's brake actuator to work according to the brake control strategy.

[0039] According to one aspect of the above technical solution, the model training module is specifically used for:

[0040] Loading, cleaning and converting the first vehicle status data, and converting the first vehicle status data into a data format that meets preset requirements;

[0041] The preprocessed first vehicle state data is used as input data and loaded into the deep learning model in batches through a data loader;

[0042] Forward propagating the first vehicle state data through the deep learning model to calculate model output data of the deep learning model;

[0043] Using a cross entropy loss function to calculate the error between the model output data and the target value;

[0044] According to the gradient of the cross entropy loss function, back propagation is performed to calculate the gradient of each parameter in the deep learning model;

[0045] Use an optimizer to update the corresponding parameters of the deep learning model according to the gradient of each parameter to obtain a target model.

[0046] The second aspect of the present invention is to provide a readable storage medium on which computer instructions are stored, and when the instructions are executed by a processor, the method described in the above technical solution is implemented.

[0047] The fourth aspect of the present invention is to provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method described in the above technical solution when executing the program.

[0048] Compared with the prior art, the brake control method, system, storage medium and device based on deep learning shown in the present invention have the following beneficial effects:

[0049] First, during the historical driving process of the vehicle, the first vehicle status data is collected through a preset sensor, and the first vehicle status data includes driving speed, braking force and acceleration. Secondly, the first vehicle status data is processed, and the first vehicle status data is imported into a pre-provided deep learning model for model training to obtain a target model. Then, during the driving process of the vehicle, the second vehicle status data is collected in real time through a preset sensor, and the second vehicle status data includes driving speed, braking force and acceleration. Finally, the second vehicle status data is imported into the target model, and a braking control strategy is output to control the vehicle's brake actuator to work according to the braking control strategy. In the present invention, the deep learning model is trained by collecting the first vehicle status data during the driving process of the vehicle, and then the second vehicle status data is imported into the target model after being collected, so that the braking control strategy can be output. The braking control according to the braking control strategy can achieve precise control of the braking process, thereby improving the user's comfort and satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the description of the embodiments in conjunction with the following drawings, in which:

[0051] Figure 1 It is a flowchart of a brake control method based on deep learning in one embodiment of the present invention;

[0052] Figure 2 This is a structural block diagram of a brake control system based on deep learning in one embodiment of the present invention. DETAILED DESCRIPTION

[0053] In order to make the purpose, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings. Several embodiments of the present invention are shown in the accompanying drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, the purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive.

[0054] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used herein in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more related listed items.

[0055] Embodiment 1

[0056] See also Figure 1The first embodiment of the present invention provides a brake control method based on deep learning, the method comprising steps S10 to S40:

[0057] Step S10: In the historical driving process of the vehicle, first vehicle status data is collected through a preset sensor, where the first vehicle status data includes driving speed, braking force and acceleration.

[0058] First of all, it should be explained that in this embodiment, the historical driving process of the vehicle can be the driving process of a vehicle driven by a specific user, that is, the driving process of a target vehicle that needs to be braked based on deep learning, or it can be the historical driving process of vehicles driven by many historical users. In this embodiment, the first vehicle state data generated during the vehicle driving process is collected by the predictive sensor, and the first vehicle state data such as the vehicle's driving speed, braking force and acceleration.

[0059] The braking force in the first vehicle data includes a plurality of braking force distribution data corresponding to a plurality of driving speed intervals, and the acceleration includes a lateral acceleration and a longitudinal acceleration during the driving process of the vehicle.

[0060] Step S20: Process the first vehicle status data, import the first vehicle status data into a pre-provided deep learning model for model training, and obtain a target model.

[0061] In this embodiment, the step of processing the first vehicle state data, importing the first vehicle state data into a pre-provided deep learning model for model training, and obtaining a target model includes:

[0062] Loading, cleaning and converting the first vehicle status data, and converting the first vehicle status data into a data format that meets preset requirements;

[0063] The preprocessed first vehicle state data is used as input data and loaded into the deep learning model in batches through a data loader;

[0064] Forward propagating the first vehicle state data through the deep learning model to calculate model output data of the deep learning model;

[0065] Using a cross entropy loss function to calculate the error between the model output data and the target value;

[0066] According to the gradient of the cross entropy loss function, back propagation is performed to calculate the gradient of each parameter in the deep learning model;

[0067] Use an optimizer to update the corresponding parameters of the deep learning model according to the gradient of each parameter to obtain a target model.

[0068] Specifically, after obtaining the first vehicle status data, the first vehicle status data will first be loaded, cleaned and converted to convert the first vehicle status data into a data format that meets the requirements of deep learning. In the process of data cleaning, missing values, abnormal values ​​or duplicate values ​​in the data are checked and processed accordingly. For example, for missing values, methods such as filling, deleting or interpolation can be selected; for abnormal values, smoothing processing can be performed or corrections can be made according to business logic.

[0069] After preprocessing the first vehicle state data, the architecture of the deep learning model will be defined, including the input layer, hidden layer, output layer, etc. The model is usually built using deep learning frameworks such as TensorFlow / Keras and PyTorch. The deep learning model is then initialized by creating a model instance and setting hyperparameters such as learning rate, batch size, number of training rounds, etc.

[0070] During the model training process, the preprocessed first vehicle state data is used as input data and loaded into the deep learning model in batches through the data loader. The input data is then forward propagated through the model to calculate the output of the model. Then, a suitable loss function (such as cross entropy loss, mean square error loss, etc.) is used to calculate the error between the model output and the target value. Then, according to the gradient of the loss function, back propagation is performed to calculate the gradient of each parameter in the model, and the optimizer (such as Adam, SGD, etc.) is used to update the model parameters according to the gradient. Finally, during the model training process, it is also necessary to record log information such as the average loss of each round and the training time in order to monitor the training process and debug the model, so as to obtain the target model.

[0071] Step S30, during the driving process of the vehicle, second vehicle status data is collected in real time through a preset sensor, and the second vehicle status data includes driving speed, braking force and acceleration.

[0072] In this embodiment, during the driving process of the vehicle, the second vehicle status data is collected in real time through a preset sensor, and the second vehicle status data includes the driving speed, braking force and acceleration, including:

[0073] During the driving process of the vehicle, the vehicle's driving speed, braking force and acceleration are collected in real time through preset sensors to obtain second vehicle status data;

[0074] Preprocessing the second vehicle status data, including filtering and denoising, and performing format conversion to obtain the second vehicle status data in JSON format;

[0075] The second vehicle status data is transmitted to a central processor or a data storage unit via a CAN bus of the vehicle.

[0076] Specifically, during the driving process of the vehicle, the pre-installed sensors will first collect data in real time and output the data in the form of electrical signals, analog signals or digital signals. For example, if the sensor outputs an analog signal, an analog-to-digital converter (ADC) is needed to convert it into a digital signal for subsequent processing.

[0077] After acquiring the second vehicle state data, the collected second vehicle state will be preprocessed, including filtering, amplification, denoising and other steps to improve the accuracy and reliability of the data. In addition, the processed data is converted into a format suitable for subsequent processing or storage, and the standard JSON format is preferably used in this embodiment. For example, the data can be packaged into a specific data packet format, or converted into a standard file format (such as CSV, JSON, etc.). Finally, the processed data is transmitted to a central processor or data storage device via the CAN bus or other wireless communications (such as Bluetooth, Wi-Fi, etc.) inside the vehicle.

[0078] Step S40, importing the second vehicle state data into the target model, and outputting a brake control strategy to control the brake actuator of the vehicle to work according to the brake control strategy.

[0079] In this embodiment, the step of importing the second vehicle state data into the target model and outputting a brake control strategy to control the brake actuator of the vehicle to work according to the brake control strategy includes:

[0080] calling the target model and continuously importing the continuously generated second vehicle state data into the target model;

[0081] Controlling the target model to output data parameters of braking timing node, braking force and braking duration according to the second vehicle state data to obtain a braking control strategy;

[0082] Performing data conversion on the brake control strategy to obtain a control signal for controlling a brake actuator;

[0083] The control signal is transmitted to the brake actuator via the CAN bus, so that the brake actuator performs corresponding braking operations according to the instructions of the control signal.

[0084] Wherein, the method further comprises:

[0085] When the brake actuator performs a corresponding brake operation according to the instruction of the control signal, the state of the brake actuator is monitored in real time by a preset brake sensor to obtain monitoring data;

[0086] The brake control strategy is adjusted in real time according to the monitoring data.

[0087] The step of adjusting the brake control strategy in real time according to the monitoring data includes:

[0088] Comparing the monitoring data with the expected effect data of the brake control strategy to determine whether there is a deviation or anomaly;

[0089] If so, evaluating the expected effect data of the current brake control strategy, identifying the parameters that need to be adjusted in the brake control strategy based on the evaluation results, and obtaining an adjustment strategy;

[0090] The brake control strategy is adjusted in real time according to the adjustment strategy.

[0091] Specifically, in this embodiment, when controlling the brake actuator to work, the target model will be called first, and the second vehicle state data will be imported into the target model, and then the target model will process and analyze the second vehicle state data according to the preset algorithm and logic. Specifically, the target model will calculate the brake control strategy based on the input data, and the brake control strategy data includes parameters such as brake timing, brake force, and brake duration.

[0092] After outputting the brake control strategy, the brake control strategy is also converted into a control signal that the brake actuator can understand, including converting the brake parameters (such as brake force) into electrical signals or air pressure signals, etc., and then transmitting the control signal to the brake actuator through the vehicle's internal communication network (such as CAN bus) or wired connection. After receiving the control signal, the brake actuator will perform the corresponding braking operation according to the instructions corresponding to the control signal, including adjusting the pressure of the brake pads, changing the flow of the brake fluid, etc.

[0093] During the braking operation, sensors or other monitoring equipment are used to monitor the state of the brake actuator and the vehicle's response in real time, including the wear of the brake pads and the temperature of the brake fluid. The brake control strategy is adjusted in real time based on the monitoring results. If the vehicle's response is not as expected, the brake parameters need to be adjusted or the deep learning model needs to be retrained.

[0094] Compared with the prior art, the brake control method based on deep learning shown in this embodiment has the following beneficial effects:

[0095] First, during the historical driving process of the vehicle, the first vehicle state data is collected through a preset sensor, and the first vehicle state data includes driving speed, braking force and acceleration. Secondly, the first vehicle state data is processed, and the first vehicle state data is imported into a pre-provided deep learning model for model training to obtain a target model. Then, during the driving process of the vehicle, the second vehicle state data is collected in real time through a preset sensor, and the second vehicle state data includes driving speed, braking force and acceleration. Finally, the second vehicle state data is imported into the target model, and a braking control strategy is output to control the vehicle's brake actuator to work according to the braking control strategy. In this embodiment, the deep learning model is trained by collecting the first vehicle state data during the driving process of the vehicle, and then the second vehicle state data is imported into the target model after being collected, so that the braking control strategy can be output. The braking control according to the braking control strategy can achieve precise control of the braking process, thereby improving the user's comfort and satisfaction.

[0096] Embodiment 2

[0097] See also Figure 2 A second embodiment of the present invention provides a brake control system based on deep learning, which is applied to the method described in the above embodiment, and the system includes:

[0098] A first data acquisition module 10 is used to collect first vehicle status data through a preset sensor during the historical driving process of the vehicle, wherein the first vehicle status data includes driving speed, braking force and acceleration;

[0099] A model training module 20, configured to process the first vehicle state data, import the first vehicle state data into a pre-provided deep learning model for model training, and obtain a target model;

[0100] The second data acquisition module 30 is used to collect the second vehicle status data in real time through the preset sensor during the driving process of the vehicle, and the second vehicle status data includes the driving speed, braking force and acceleration;

[0101] The strategy output module 40 is used to import the second vehicle state data into the target model and output a brake control strategy to control the brake actuator of the vehicle to work according to the brake control strategy.

[0102] Wherein, the model training module 20 is specifically used for:

[0103] Loading, cleaning and converting the first vehicle status data, and converting the first vehicle status data into a data format that meets preset requirements;

[0104] The preprocessed first vehicle state data is used as input data and loaded into the deep learning model in batches through a data loader;

[0105] Forward propagating the first vehicle state data through the deep learning model to calculate model output data of the deep learning model;

[0106] Using a cross entropy loss function to calculate the error between the model output data and the target value;

[0107] According to the gradient of the cross entropy loss function, back propagation is performed to calculate the gradient of each parameter in the deep learning model;

[0108] Use an optimizer to update the corresponding parameters of the deep learning model according to the gradient of each parameter to obtain a target model.

[0109] Compared with the prior art, the brake control system based on deep learning shown in this embodiment has the following beneficial effects:

[0110] First, during the historical driving process of the vehicle, the first vehicle state data is collected through a preset sensor, and the first vehicle state data includes driving speed, braking force and acceleration. Secondly, the first vehicle state data is processed, and the first vehicle state data is imported into a pre-provided deep learning model for model training to obtain a target model. Then, during the driving process of the vehicle, the second vehicle state data is collected in real time through a preset sensor, and the second vehicle state data includes driving speed, braking force and acceleration. Finally, the second vehicle state data is imported into the target model, and a braking control strategy is output to control the vehicle's brake actuator to work according to the braking control strategy. In this embodiment, the deep learning model is trained by collecting the first vehicle state data during the driving process of the vehicle, and then the second vehicle state data is imported into the target model after being collected, so that the braking control strategy can be output. The braking control according to the braking control strategy can achieve precise control of the braking process, thereby improving the user's comfort and satisfaction.

[0111] Embodiment 3

[0112] A third embodiment of the present invention provides a readable storage medium on which computer instructions are stored. When the instructions are executed by a processor, the method described in the above embodiment is implemented.

[0113] Embodiment 4

[0114] A fourth embodiment of the present invention provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method described in the above embodiments when executing the program.

[0115] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0116] The above-mentioned embodiments only express several implementation methods of the present invention, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for those of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. A brake control method based on deep learning, characterized in that: The method comprises: During the historical driving process of the vehicle, first vehicle status data is collected through a preset sensor, wherein the first vehicle status data includes driving speed, braking force and acceleration; Performing data processing on the first vehicle state data, importing the first vehicle state data into a pre-provided deep learning model for model training, and obtaining a target model; During the driving process of the vehicle, the second vehicle status data is collected in real time through a preset sensor, and the second vehicle status data includes driving speed, braking force and acceleration; The second vehicle state data is imported into the target model, and a brake control strategy is output to control the brake actuator of the vehicle to work according to the brake control strategy.

2. The brake control method based on deep learning according to claim 1, characterized in that: The step of processing the first vehicle status data, importing the first vehicle status data into a pre-provided deep learning model for model training, and obtaining a target model includes: Loading, cleaning and converting the first vehicle status data, and converting the first vehicle status data into a data format that meets preset requirements; The preprocessed first vehicle state data is used as input data and loaded into the deep learning model in batches through a data loader; Forward propagating the first vehicle state data through the deep learning model to calculate model output data of the deep learning model; Using a cross entropy loss function to calculate the error between the model output data and the target value; According to the gradient of the cross entropy loss function, back propagation is performed to calculate the gradient of each parameter in the deep learning model; Use an optimizer to update the corresponding parameters of the deep learning model according to the gradient of each parameter to obtain a target model.

3. The brake control method based on deep learning according to claim 1, characterized in that: During the driving process of the vehicle, a step of collecting second vehicle status data in real time through a preset sensor, wherein the second vehicle status data includes driving speed, braking force and acceleration, comprises: During the driving process of the vehicle, the vehicle's driving speed, braking force and acceleration are collected in real time through preset sensors to obtain second vehicle status data; Preprocessing the second vehicle status data, including filtering and denoising, and performing format conversion to obtain the second vehicle status data in JSON format; The second vehicle status data is transmitted to a central processor or a data storage unit via a CAN bus of the vehicle.

4. The brake control method based on deep learning according to any one of claims 1 to 3, characterized in that: The step of importing the second vehicle state data into the target model and outputting a brake control strategy to control the brake actuator of the vehicle to work according to the brake control strategy includes: calling the target model and continuously importing the continuously generated second vehicle state data into the target model; Controlling the target model to output data parameters of braking timing node, braking force and braking duration according to the second vehicle state data to obtain a braking control strategy; Performing data conversion on the brake control strategy to obtain a control signal for controlling a brake actuator; The control signal is transmitted to the brake actuator via the CAN bus, so that the brake actuator performs corresponding braking operations according to the instructions of the control signal.

5. The brake control method based on deep learning according to claim 4 is characterized in that: The method further comprises: When the brake actuator performs a corresponding brake operation according to the instruction of the control signal, the state of the brake actuator is monitored in real time by a preset brake sensor to obtain monitoring data; The brake control strategy is adjusted in real time according to the monitoring data.

6. The brake control method based on deep learning according to claim 5, characterized in that: The step of adjusting the brake control strategy in real time according to the monitoring data comprises: Comparing the monitoring data with the expected effect data of the brake control strategy to determine whether there is a deviation or anomaly; If so, evaluating the expected effect data of the current brake control strategy, identifying the parameters that need to be adjusted in the brake control strategy based on the evaluation results, and obtaining an adjustment strategy; The brake control strategy is adjusted in real time according to the adjustment strategy.

7. A brake control system based on deep learning, characterized in that: The method applied to any one of claims 1 to 6, wherein the system comprises: A first data acquisition module, used to collect first vehicle status data through a preset sensor during the historical driving process of the vehicle, wherein the first vehicle status data includes driving speed, braking force and acceleration; A model training module, used to process the first vehicle state data, import the first vehicle state data into a pre-provided deep learning model for model training, and obtain a target model; A second data acquisition module, used to collect second vehicle status data in real time through a preset sensor during the driving process of the vehicle, wherein the second vehicle status data includes driving speed, braking force and acceleration; The strategy output module is used to import the second vehicle state data into the target model and output a brake control strategy to control the vehicle's brake actuator to work according to the brake control strategy.

8. The deep learning-based brake control system according to claim 7, characterized in that: The model training module is specifically used for: Loading, cleaning and converting the first vehicle status data, and converting the first vehicle status data into a data format that meets preset requirements; The preprocessed first vehicle state data is used as input data and loaded into the deep learning model in batches through a data loader; Forward-propagating the first vehicle state data through the deep learning model to calculate model output data of the deep learning model; Using a cross entropy loss function to calculate the error between the model output data and the target value; Perform back propagation according to the gradient of the cross entropy loss function to calculate the gradient of each parameter in the deep learning model; Use an optimizer to update the corresponding parameters of the deep learning model according to the gradient of each parameter to obtain a target model.

9. A readable storage medium having computer instructions stored thereon, characterized in that: When the instruction is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

10. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the method according to any one of claims 1 to 6 is implemented.

Images