CN118372788A - Method, device, controller, vehicle and medium for comfortable braking of vehicle - Google Patents

Abstract

Embodiments of the present disclosure relate to methods, devices, controllers, vehicles, and media for comfortable braking of a vehicle. The method includes updating a plurality of model parameters for the vehicle, such as automatically acquiring updated parameters of the artificial intelligence model. The method also includes determining a comfort braking force for the vehicle based on a plurality of vehicle parameters and a plurality of model parameters associated with the vehicle. In addition, the method includes controlling a comfortable braking of the vehicle based on the comfortable braking force. Therefore, according to the scheme provided by the embodiment of the disclosure, the model parameters can be automatically updated, and the comfortable braking force is determined through the updated model parameters and the updated vehicle parameters.

Description

Method, device, controller, vehicle and medium for comfortable braking of vehicle

Technical Field

Embodiments of the present disclosure relate to the field of vehicle control, and more particularly, to methods, devices, controllers, vehicles, and media for comfortable braking of a vehicle.

Background

The braking experience is very important during driving of the vehicle. The vehicle is inevitably subjected to the vehicle body pause during braking or accelerating, particularly in the process of starting braking until stopping, the vehicle can experience a plurality of pauses, which can cause discomfort to drivers and passengers and even cause potential safety hazards.

The powertrain or suspension system of the vehicle is typically optimized to mitigate this jerk. But the comfortable braking technology is utilized to optimize the braking process of the vehicle, and smooth parking is realized by intelligently adjusting the braking force, so that the jerk phenomenon can be effectively reduced. The brake system with the comfortable brake technology not only creates good driving experience for a driver, but also improves driving safety.

Disclosure of Invention

Embodiments of the present disclosure relate to methods, devices, controllers, vehicles, and media for comfortable braking of a vehicle.

According to a first aspect of the present disclosure, a method for comfortable braking of a vehicle is provided. The method includes updating a plurality of model parameters for the vehicle. The method further includes determining a comfort braking force for the vehicle based on a plurality of vehicle parameters related to the vehicle and the plurality of model parameters. In addition, the method includes controlling comfortable braking of the vehicle based on the comfortable braking force.

According to a second aspect of the present disclosure, an apparatus for comfortable braking of a vehicle is provided. The apparatus includes a model parameter updating unit configured to update a plurality of model parameters for the vehicle. The apparatus further includes a comfort brake determination unit configured to determine a comfort braking force for the vehicle based on a plurality of vehicle parameters related to the vehicle and the plurality of model parameters. Furthermore, the apparatus comprises a comfort brake control unit configured to control a comfort brake of the vehicle based on the comfort brake force.

According to a third aspect of the present 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 that, when executed by the at least one processor, cause the controller to perform the steps of the method in the first aspect of the present disclosure.

According to a fourth aspect of the present disclosure there is provided a vehicle comprising the controller of the third aspect of the present disclosure.

According to a fifth aspect of the present disclosure, a machine-readable storage medium is provided. The machine-readable storage medium has stored thereon machine-executable instructions which are executed by a processor to implement the steps of the method in the first aspect of the present disclosure.

Drawings

The foregoing and other objects, features and advantages of the disclosure will be apparent from the following more particular descriptions of exemplary embodiments of the disclosure as illustrated in the accompanying drawings wherein like reference numbers generally represent like parts throughout the exemplary embodiments of the disclosure.

FIG. 1 illustrates a schematic diagram of an example environment in which devices and/or methods may be implemented, according to embodiments of the present disclosure;

FIG. 2 illustrates a flow chart of a method for comfortable braking of a vehicle according to an embodiment of the present disclosure;

FIG. 3 illustrates a flow chart of a process of triggering model training in accordance with an embodiment of the present disclosure;

FIG. 4 illustrates a flowchart of a process of training and deploying a comfort braking model in accordance with an embodiment of the present disclosure;

FIG. 5A illustrates a schematic diagram of a process of training a linear regression model according to an embodiment of the present disclosure;

FIG. 5B illustrates a schematic diagram of a process of predicting using a linear regression model according to an embodiment of the present disclosure;

FIG. 6 illustrates a schematic diagram of a process of online training and prediction of a comfort braking model in accordance with an embodiment of the present disclosure;

FIG. 7 illustrates a schematic diagram of a model triggering and deployment procedure, according to an embodiment of the present disclosure;

FIG. 8 illustrates a schematic view of an apparatus for comfortable braking of a vehicle in accordance with an embodiment of the present disclosure; and

Fig. 9 illustrates a schematic block diagram of an example device suitable for use in practicing embodiments of the disclosure.

Detailed Description

Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While certain embodiments of the present disclosure have been shown in the accompanying drawings, it is to be understood that the present disclosure may be embodied in various forms and should not be construed as limited to the embodiments set forth herein, but are provided to provide a more thorough and complete understanding of the present disclosure.

As described above, during braking, comfortable braking and stopping can bring a good braking experience to the driver, and avoid the phenomenon that the vehicle is in a bump during braking, especially when stopping immediately. In the conventional comfortable braking technology, when the comfortable braking force for braking the vehicle is determined, the comfortable braking force is often calculated according to the calibration parameters, and the calibration parameters are not adjusted, so the calibration parameters for calculating the comfortable braking force are often fixed. However, due to long-short-period changing factors of the vehicle and parts thereof and some driving environment factors, parameters related to the factors are not considered in the comfort braking force calculated through the fixed calibration parameters, so that the comfort braking force cannot be self-adaptive along with the change of the execution environment, and further, the braking experience brought to drivers cannot be in an optimal state all the time. For example, the vehicle itself and its components may wear due to long-term driving, possibly resulting in a calculated comfort braking force based on fixed calibration parameters that is less or greater than the actually required comfort braking force. In addition, some driving environment factors such as a change in wind resistance caused by wind speed, a change in friction coefficient of a brake disc caused by air temperature, a change in speed at which pressure is generated by brake fluid, and the like may cause that a comfortable braking force calculated according to a fixed calibration parameter cannot achieve optimal comfortable braking.

To this end, embodiments of the present disclosure propose a scheme for comfortable braking of a vehicle that first updates a plurality of model parameters of the vehicle, and determines a comfortable braking force of the vehicle from the plurality of model parameters and the plurality of vehicle parameters, and then controls the comfortable braking of the vehicle using the comfortable braking force. Therefore, according to the scheme for the comfortable braking of the vehicle, the plurality of model parameters can be updated, and the comfortable braking force is determined through the updated plurality of model parameters and the plurality of vehicle parameters.

Embodiments of the present disclosure will be described in further detail below with reference to the drawings, wherein FIG. 1 illustrates an example environment 100 in which the devices and/or methods of embodiments of the present disclosure may be implemented.

As shown in FIG. 1, example environment 100 includes vehicles 110-1 through 110-N (hereinafter referred to as vehicles 110 individually or collectively). Vehicle 110 includes sensors 112, comfort brake model 114, memory 116, and controller 118. The sensors 112 may include, but are not limited to, one or more of the following: a brake pedal sensor, an acceleration sensor, a wheel speed sensor, a gyroscope sensor, a tire pressure detection sensor, a suspension height sensor, a door sensor, a seat sensor, a camera, a radar sensor, a temperature sensor, a humidity sensor, and the like. A plurality of vehicle parameters associated with vehicle 110 may be acquired by sensor 112. In some embodiments, the vehicle parameters include, but are not limited to, a current speed of the vehicle, a deceleration of the vehicle, a weight of the vehicle, a brake disc temperature of the vehicle, a brake disc friction coefficient of the vehicle, or a brake disc wear coefficient of the vehicle, and the like.

Further, the opening degree of the brake pedal may be acquired by a brake pedal sensor to determine the braking force requested by the driver. The acceleration sensor is typically mounted on the chassis of the vehicle 110, and measures acceleration by detecting movement of the chassis, and deceleration of the vehicle (the same magnitude as the acceleration, opposite in direction) is an important parameter for determining the comfort braking force when the vehicle is braked, and the larger the deceleration, the larger the comfort braking force. Inertial sensors can detect acceleration, speed, and direction changes of the vehicle. The wheel speed sensor can acquire the rotational speed of each wheel of the vehicle 110, and the influence of the wheel speed is great in determining the comfortable braking force for the vehicle 110.

In addition, the gyro sensor may acquire the rotational speed of the vehicle 110, which is also a factor to be considered in calculating the comfort braking force. In addition, a camera and radar sensor may be used to determine the distance of the vehicle 110 from the obstacle to determine if the comfort brake may be turned on and to calculate the magnitude of the comfort brake force based on the distance. The temperature and humidity sensors may acquire the temperature and humidity of critical components of the vehicle 110, such as the brake fluid and the brake disc, which are also important for determining a comfortable braking force. It should be understood that various sensors and corresponding vehicle parameters are listed herein, but the present disclosure is not limited to the number and type of sensors and the number and type of vehicle parameters, and embodiments of the present disclosure may utilize more or fewer vehicle parameters to determine a comfort braking force.

With continued reference to FIG. 1, the vehicle 110 also includes a comfort braking model 114. The comfort brake model 114 may include, but is not limited to, a machine learning model, a deep learning model, or an artificial intelligence model, among others. As previously described, utilizing the comfort braking model 114 and corresponding plurality of vehicle parameters to calculate the comfort braking force can take into account various factors associated with the vehicle 110, and can result in a more accurate comfort braking force as compared to conventional comfort braking forces calculated from fixed calibration parameters, as the comfort braking model can better model the comfort braking force. Furthermore, better braking experience can be brought to a driver by using more accurate comfortable braking force, and negative influence brought by change of the execution environment of the comfortable braking force is avoided.

In addition, the vehicle includes a memory 116, and a plurality of model parameters associated with the comfort braking model may be stored in the memory 116. For example, when the comfort braking model is a linear regression model, the weight vector W (i.e., model parameters) for each feature of the comfort braking model (i.e., vehicle parameters) may be stored in the memory 116. Furthermore, when the comfort brake model is a neural network model, the weight matrix W (i.e., model parameters) for each layer of the comfort brake model may be stored in the memory 116. In addition, a plurality of model parameters may be updated to better accommodate changes in factors associated with vehicle 110. For example, the brake disc wears after long driving of the vehicle, and thus both the brake disc friction coefficient and the brake disc wear coefficient change, thereby affecting the calculation of the comfortable braking force, and thus a plurality of model parameters need to be updated to cope with such changes. In some embodiments, the comfort brake model 114 may be trained on the vehicle 110 to update a plurality of model parameters and stored in the memory 116. In some embodiments, the comfort brake model 114 is only model architecture information, while the model is trained on the cloud server 140 and only model architecture information is saved as the comfort brake model 114, while the supporting model parameters are updated into the memory 116. Because only the comfort braking model 114 and model parameters on the memory 116 need be utilized to determine the comfort braking force, the amount of computation on the vehicle 110 may be reduced. It should be appreciated that the storage of a plurality of model parameters on the memory 116 is merely exemplary and that model parameters may be stored in other components of the vehicle.

With continued reference to FIG. 1, the vehicle 110 further includes a controller 118, and the controller 118 may receive the plurality of vehicle parameters acquired by the sensors 112 and calculate a comfort braking force from the comfort braking model 114 and the plurality of model parameters stored on the memory 116. In addition, the controller 118 may obtain a comfortable braking force to control the comfortable braking process of the vehicle 110. For example, a comfortable braking force is applied to the brake actuators 120 of the vehicle 110 to control a comfortable braking process of the vehicle 110, it being understood that the brake actuators may be mounted on each wheel, and further, the brake actuators 120 are shown as an example only. Further, the example environment 100 shows the comfort brake model 114 and the memory 116 disposed separately from the controller 118, but the comfort brake model 114 and/or the memory 116 may be disposed in the controller 118, which is not limited in this disclosure.

In addition, the example environment 100 also includes a cloud server 140. In some embodiments, comfort brake model 144 may be deployed on cloud server 140 and training data is obtained from vehicle 110 over network 130 to train comfort brake model 144, and architectural information of comfort brake model 144 (e.g., comfort brake model 114) may be saved on vehicle 110, with only the train updated plurality of model parameters sent to vehicle 110 and saved in memory 116. Further, multiple comfort braking models may be deployed on the cloud server 140. For example, a comfort braking model may be deployed for each vehicle, or a comfort braking model may be deployed for each model of vehicle, so that a plurality of comfort braking models may be deployed on the cloud server 140, so that the comfort braking force may be more accurately calculated since the comfort braking model is specific to the vehicle and/or model. In addition, a comfort braking model can be deployed for all vehicles, so that the comfort braking model can be trained by training data of all vehicles, and therefore the comfort braking model can be trained more accurately, and the comfort braking force can be calculated more accurately.

In addition, a plurality of databases 142 may be included on cloud server 140, and databases 142 may store training data related to vehicle 110 to train comfort brake model 144 on cloud server 140 with the training data. In some embodiments, database 142 may also hold historical training data for vehicle 110, and may be trained based on the historical training data and current training data each time comfort brake model 144 is trained, so that comfort brake model 144 may be trained more fully.

With continued reference to fig. 1, the example environment 100 also includes a network 130. In an embodiment of the present disclosure, vehicle 110 and cloud server 140 enable wireless transmission of data over network 130. For example, vehicle 110 may transmit training data, which may be a plurality of vehicle parameters of vehicle 110, such as speed, acceleration, etc., over network 130 to cloud server 140 that is required to train comfort brake model 144. In addition, a plurality of model parameters of the trained comfort brake model 144 may also be transmitted to the vehicle 110 over the network 130 and stored in the memory 116 to effect updating of the model parameters. It should be appreciated that cloud server 140 and network 130 are shown by way of example only, as the comfort brake model may be trained on vehicle 110, model training and model parameter updating may not require cloud server 140 and network 130 when the comfort brake model is trained on vehicle 110 only.

An example environment 100 in which embodiments of the present disclosure can be implemented is described above in connection with fig. 1. A flowchart of a method 200 for comfortable braking of a vehicle according to an embodiment of the present disclosure is described below in connection with fig. 2.

As shown in fig. 2, at block 202, a plurality of model parameters for a vehicle may be updated. For example, referring to fig. 1, a plurality of model parameters of the vehicle 100 may be saved on the memory 116 and may be updated on the vehicle 100 by the controller 118 of the vehicle or directed to the cloud server 140 to update the plurality of model parameters.

At block 204, a comfort braking force for the vehicle may be determined based on a plurality of vehicle parameters and a plurality of model parameters related to the vehicle. For example, in connection with FIG. 1, a plurality of vehicle parameters associated with vehicle 110 may be acquired by sensor 112 and utilized by vehicle controller 118 to determine a comfort braking force for vehicle 100 using the plurality of vehicle parameters and the plurality of model parameters.

At block 206, comfort braking of the vehicle may be controlled based on the comfort braking force. For example, referring to fig. 1, a comfort braking force may be determined by the controller 118 and comfort braking of the vehicle may be controlled according to the magnitude of the comfort braking force.

Therefore, by the method 200 provided by the embodiment of the present disclosure, a plurality of model parameters can be updated, and further, a comfortable braking force is determined by the updated plurality of model parameters and a plurality of vehicle parameters, and compared with the existing method, since the plurality of model parameters can be automatically updated instead of using fixed calibration parameters, the comfortable braking force can be more accurately calculated, thereby providing a better comfortable braking experience for a driver when the vehicle brakes, and maintaining such an optimal state in the whole life cycle of a comfortable braking product.

Fig. 3 illustrates a flow chart of a process 300 of triggering model training in accordance with an embodiment of the present disclosure. As shown in fig. 3, at block 302, a speed and a deceleration of a vehicle may be obtained. As described above, the current speed of the vehicle may be acquired by the wheel speed sensor, for example, the wheel speed is converted into a linear speed, i.e., the current speed of the vehicle, according to the wheel speed and the wheel diameter. Further, in some embodiments, the current speed of the vehicle may also be obtained by a radar sensor, which is not limiting to the present disclosure.

At block 304, a target deceleration of the vehicle may be determined based on a speed of the vehicle. In some embodiments, a golden deceleration curve may be utilized to determine a target deceleration of the vehicle. For example, the golden deceleration curve may be a calibrated deceleration curve indicating a corresponding target deceleration at a certain vehicle speed, and the braking procedure corresponding to the golden deceleration curve may enable comfortable braking expected by the driver. Further, in some embodiments, the target deceleration may be determined using a target deceleration model. For example, the corresponding target deceleration may be determined by a target deceleration model based on the current vehicle speed.

At block 306, a difference between the actual deceleration value and the target deceleration value may be determined. For example, at the target deceleration a ₁ corresponding to the current vehicle speed V, the current actual deceleration is a ₂, and thus the difference between the target deceleration a ₁ and the actual deceleration a ₂ can be determined by comparing them.

At block 308, it may be determined whether the difference between the actual deceleration value and the target deceleration value is greater than a difference threshold. For example, the difference value is d= (a ₁–a₂), and the difference threshold may be a preset threshold t, which may be a difference threshold empirically set by an engineer related to the field.

At block 308, if the difference value d is greater than the threshold t, then proceed to block 310 to trigger training of the comfort brake model and update the plurality of model parameters accordingly. As depicted in fig. 1, training of the comfort brake model may be performed on cloud processor 140 and after training is completed, the trained plurality of model parameters are transmitted to vehicle 110 to enable updating of the plurality of model parameters. If the difference value d is less than or equal to the threshold value t, proceeding to block 312, training of the comfort braking model is not triggered, and accordingly, updating of the plurality of model parameters is not required. Since the comfort braking force generated by the current comfort braking model can meet the comfort braking requirement if the difference value d is less than or equal to the threshold value t, no update is required.

Further, in some embodiments, it may also be determined whether to trigger training of the comfort braking force model by determining a difference between the actual comfort braking force and the target comfort braking force during the comfort braking of the vehicle. For example, if the actual comfort braking force is greater than the target comfort braking force and the difference between the two is greater than the threshold value, it means that the vehicle may not achieve optimal comfort braking during braking, and the actual comfort braking force may be too great, which may still cause the vehicle to feel a jerk, resulting in poor braking experience for the driver. Further, if the actual comfortable braking force is smaller than the target comfortable braking force and the difference therebetween is larger than the threshold value, this means that the vehicle may be braked too slowly at the time of braking, the braking distance is prolonged, and there is a slight feeling of slipping when the driver is sensitive to a change in the deceleration of the vehicle. Thus, in all of the above cases, it is necessary to trigger training of the comfort brake model and update the model parameters on the vehicle.

In some embodiments, the magnitude of the bump of the vehicle suspension may be obtained by a sensor and a determination may be made as to whether to perform a model update based on the magnitude of the bump. By the suspension bump width, whether the comfortable braking force is proper or not can be judged. For example, if the amplitude of the suspension jerk is too large, this means that the comfort braking force is too great and that comfort braking cannot be achieved, thus requiring training of the comfort braking model to be triggered and model parameters on the vehicle to be updated. If the amplitude of the suspension jerk is small, this means that the comfort braking force is appropriate and that a comfort braking can be achieved, so that there is no need to adjust the comfort braking model.

Thus, through process 300 of embodiments of the present disclosure, it may be determined whether to trigger training of the comfort brake model and updating of the model parameters through one or more vehicle parameters. The conventional comfort braking technique utilizes a calibrated comfort braking force and is not automatically updated, and the embodiments of the present disclosure can automatically update the comfort braking model through the above-described process, and thus can be timely adjusted when the vehicle itself and its components or vehicle environmental factors change, thereby enabling more accurate determination of the comfort braking force. Further, because the process 300 of embodiments of the present disclosure is updated by conditional triggers, rather than timed updates (e.g., daily, weekly, monthly, etc.), unnecessary model training and computation may be reduced, thereby saving costs for model training and data transmission.

Fig. 4 illustrates a flowchart of a process 400 of training and deploying a comfort braking model in accordance with an embodiment of the present disclosure. As shown in fig. 4, at block 402, a plurality of vehicle parameters of a vehicle are acquired. For example, in some embodiments, a current speed and a current deceleration of the vehicle may be obtained. In some embodiments, the temperature and humidity of the vehicle's brake disc may also be obtained. In addition, in some embodiments, the wind speed of the environment in which the vehicle is located, whether it is raining, and so on, may also be obtained. Embodiments of the present disclosure may include the above plurality of parameters, but are not limited to the above parameters, and may include any vehicle-itself parameters (e.g., speed, deceleration) related to vehicle braking, as well as vehicle environmental parameters (e.g., wind speed, rainfall).

At block 404, the driver requested braking force is obtained. In some embodiments, the requested braking force may be determined by an opening degree of a brake pedal. Thus, a relationship between the requested braking force and a plurality of vehicle parameters may be established, i.e. modeled with a comfortable braking model. At block 406, a plurality of vehicle parameters and a requested braking force may be uploaded to a cloud server. For example, referring to fig. 1, a plurality of vehicle parameters and requested braking forces may be uploaded to cloud server 140 via network 130.

At block 408, a Shu Zhidong model may be trained on the cloud server. A comfort braking model is trained by a plurality of vehicle parameters and a requested braking force, and a relationship between the requested braking force and the plurality of vehicle parameters is established. In addition, during the training process of the comfortable braking model, the model parameters of the comfortable braking model are continuously and iteratively updated.

At block 410, the updated model parameters may be sent to the vehicle. For example, referring to fig. 1, comfort brake model 144 may be trained on cloud server 140 and, after training is completed, model parameters of comfort brake model 144 are sent to vehicle 110 over network 130. The comfort braking force may then be calculated by a simple comfort braking model deployed on the vehicle 110, namely the comfort braking model 114. It should be appreciated that the comfort brake model 114 on the vehicle 100 has the same model architecture as the comfort brake model 144 on the cloud 140, but that the comfort brake model 114 only needs to maintain comfort brake model 144 architecture information, and does not need to train the comfort brake model 144. In addition, when the comfort brake model 144 incorporates the updated plurality of model parameters, the generated comfort brake force is the same magnitude as the comfort brake force generated by the comfort brake model 144.

Further, in some embodiments, the comfort braking model may be updated on the vehicle. For example, referring to FIG. 1, a training update may be performed on the comfort brake model 114. Therefore, operations such as uploading training data or sending model parameters can be avoided, and the model and the parameters can be updated only by the vehicle.

Fig. 5A illustrates a schematic diagram of a process 500A of training a linear regression model according to an embodiment of the present disclosure. In fig. 5A, the training may be performed using a linear regression model as the comfort braking model, which is simple and intuitive, computationally efficient, and interpretable, it being understood that the linear regression is shown for only one embodiment of the comfort braking model, which is not limited to any machine learning model, deep learning model, or artificial intelligence model. In addition, the linear regression model described below is a comfortable braking model, and is described using the linear regression model for convenience of understanding.

At block 502, a model expression of a linear regression model may be constructed. For example, F _Total＝F_Mass+F_wind+F_Other, the drag force of the braking force in relation to the weight of the vehicle, the wind drag force, and other drag force configurations can be determined. Thus, we can construct a model expression of the linear regression model as shown in equation (1):

F_p ＝ W*x＝W₀ + W₁*x₁+ W₂*x₂ (1)

The vector W is a model parameter of the linear regression model, and includes a plurality of model parameters, i.e., W ₀、W₁ and W ₂. It should be understood that the three parameters shown herein are merely examples and that in practice there may be more or fewer parameters. x ₁ represents the actual deceleration of the vehicle, and W ₁ corresponds to a factor related to the vehicle weight. x ₂ square of current vehicle speed, W ₂ corresponds to wind resistance related factors, W ₀ corresponds to the influence of other factors, and the wind resistance related factors are uniformly included in W _0.

At block 504, a predicted comfort braking force may be determined based on a model expression of the linear regression model. For example, using equation (1), initialized W ₀、W₁ and W ₂, and deceleration x ₁ and square x ₂ of the current vehicle speed, the predicted comfort braking force can be calculated.

At block 506, a loss function value may be calculated. For example, the loss function value may be determined by equation (2) as follows:

Loss ＝ 1/2 * (W*x - F_actual) (2)

Wherein F _actual is the actual comfort braking force, from which the Loss function Loss can be calculated.

At block 508, a plurality of model parameters may be updated by a gradient descent method. For example, the gradient value may be determined by the following formula (3), as follows:

G ＝ (W*x - F_actual) * x (3)

Where G is the gradient value, which is obtained by deriving the Loss function Loss. After obtaining the gradient value G, a plurality of model parameters may be updated using equation (4), as follows:

W’ ＝ W – lr*G (4)

where W' is the updated model parameter vector and lr is the learning rate. Thereby, an iterative update of the model parameters can be achieved.

At block 510, a determination may be made as to whether the model update is complete. In some embodiments, whether model training is complete may be determined by whether the change in model parameters is less than a threshold. In one embodiment, a determination may be made as to whether the model loss value is small enough to determine whether model training is complete. If it is determined that the model update is not complete, then the flow proceeds to block 504 where iterative training continues. If a determination is made that the model update is complete, then proceed to block 512 to complete model training.

Thus, through the process 500A provided by the embodiments of the present disclosure, a linear regression model may be used as a comfortable braking model for training, where the linear regression model is simple and intuitive, efficient in computation, and highly interpretable, while facilitating training iterations and deployments, as just an example.

Fig. 5B illustrates a schematic diagram of a process 500B for predicting using a linear regression model according to an embodiment of the present disclosure. At block 520, updated model parameters may be obtained. In conjunction with fig. 1 and 5A, model parameters may be saved in the memory 116 and the plurality of model parameters for the linear regression model is w= (W ₀,W₁,W₂).

At block 522, a plurality of vehicle parameters of the vehicle may be acquired. As described in connection with fig. 5A, the current vehicle speed V of the vehicle may be obtained and the square x ₂ of the vehicle speed, i.e., x ₂＝V², may be derived from the vehicle speed V. At block 524, a target deceleration may be determined based on the vehicle speed. As shown in fig. 3, the golden deceleration curve and the current vehicle speed may be used to determine the target deceleration, the golden deceleration curve may be a calibrated deceleration curve, the corresponding target deceleration at a certain vehicle speed is indicated, and the braking process corresponding to the golden deceleration curve may implement comfortable braking. Thus, the target deceleration x ₁ can be determined.

At block 526, a target comfort braking force may be determined. By combining the above formula (1), the target comfort braking force can be calculated in the case where the respective parameters are known to be determined. For example, F _p＝W*x＝W₀+W₁*x₁+W₂*x₂, where each model parameter and vehicle parameter have been acquired, a target comfort braking force may be determined and the vehicle may be braked comfortably in accordance with the target comfort braking force.

Fig. 6 illustrates a schematic diagram of a process 600 of online training and prediction of a comfort braking model in accordance with an embodiment of the present disclosure. In describing the embodiment of fig. 6, still referring to fig. 5A and 5B, a linear regression model is used as the comfort braking model, it being understood that the comfort braking in embodiments of the present disclosure is not limited to the linear regression model. As shown in fig. 6, at block 602, a plurality of vehicle parameters of a vehicle are acquired during a braking cycle of the vehicle. For example, the current vehicle speed and deceleration of the vehicle may be acquired. As described in fig. 5A and 5B, the vehicle speed square may be determined from the vehicle speed, i.e., the vehicle speed square 604 (i.e., the square of the vehicle speed) may be determined, and the golden deceleration curve and the current vehicle speed may be used to determine the target deceleration 606. The target comfort braking force 610 may then be determined using a comfort braking model 608 deployed on the vehicle, as depicted in FIG. 5B.

With continued reference to fig. 6, the comfort braking of the vehicle 612 may be controlled based on the target comfort braking force 610. Further, the broken line 614 may be preceded by a braking event at the time of the vehicle t, and the broken line 614 may be followed by a braking event at the time of the vehicle t+1. In addition, the current vehicle speed 616, current deceleration 618, and target braking force 620 of the vehicle may be obtained, at block 622, to determine whether to trigger training of the comfort braking model, as described in the process 300 of triggering model training of FIG. 3. In response to triggering the training of the comfort brake model, training of the comfort brake model occurs at 624. Further, the braking process at time t+1 may continue to acquire the speed 626 and deceleration 628 of the vehicle, and a determination may be made at block 630 as to whether to trigger training of the comfort braking model based on the speed 626 and deceleration 628. In response to triggering the training of the comfort brake model, training of the comfort brake model continues at 624.

Fig. 7 illustrates a schematic diagram of a model triggering and deployment procedure 700, according to an embodiment of the present disclosure. As shown in fig. 7, at block 702, it is determined whether to trigger training of a comfort braking model based on one or more of a plurality of vehicle parameters. For example, as previously described, it may be determined whether to trigger training of the comfort brake model by comparing the target deceleration with the actual deceleration. At block 704, brake behavior data of a driver is received. In some embodiments, the braking performance data may be used to determine whether the braking process is smooth and if the braking process is not smooth enough, the data for the braking process may negatively impact the model, such as the model may learn an unstable braking process, so that the data associated with the braking process may be discarded when the braking process is not smooth enough.

At block 706, vehicle state data may be received. For example, activation and intervention of dynamic controllers on the vehicle, the detected vehicle state data may not be accurate enough, and thus the corresponding data needs to be discarded. At block 708, road condition data may be received. For example, in an extreme road such as a desert, a comfortable braking force required by a vehicle is completely different from that of a regular road, so that when it is judged from road condition data that the vehicle is on an irregular road, the corresponding training data can be discarded. At block 710, based on each of the brake behavior data, the vehicle state data, and the road condition data satisfying the predetermined condition, the flow proceeds to block 712. As described above, if any one of the brake behavior data, the vehicle state data, and the road condition data does not satisfy the predetermined condition, the corresponding training data needs to be discarded without triggering the model training. At block 712, training of a comfort braking model is performed. At block 714, a trained comfort braking model is deployed on the vehicle.

Thus, by the process 700 provided by embodiments of the present disclosure, it may be determined that training of the comfort brake model is not triggered under certain abnormal conditions, avoiding abnormal data from affecting the effectiveness of the comfort brake model.

Fig. 8 illustrates a schematic diagram of an apparatus 800 for comfortable braking of a vehicle according to an embodiment of the present disclosure. The apparatus 800 includes a model parameter updating unit 802, a comfort brake determining unit 804, and a comfort brake determining unit 806. The model parameter updating unit 802 is configured to update a plurality of model parameters for the vehicle. The comfort brake determination unit 804 is configured to determine a comfort braking force for the vehicle based on a plurality of vehicle parameters related to the vehicle and the plurality of model parameters. Further, the comfort brake determining unit 806 is configured to control a comfort brake of the vehicle based on the comfort brake force.

In some embodiments, wherein the model parameter updating unit 802 comprises: a parameter update determination unit configured to determine whether to trigger an update of the plurality of model parameters based on one or more of the plurality of vehicle parameters; and a model parameter second updating unit that updates a plurality of model parameters for the vehicle in response to determining that the updating is triggered.

In some embodiments, wherein the model parameter second updating unit comprises: a target deceleration determination unit that determines a target deceleration of the vehicle based on a vehicle speed among the plurality of vehicle parameters; a deceleration difference determination unit configured to determine a deceleration difference between a vehicle deceleration and a target deceleration among the vehicle parameters; and a model update triggering unit configured to trigger updating of the plurality of model parameters in response to the deceleration difference being greater than a difference threshold.

In some embodiments, wherein the model update triggering unit comprises: a training data acquisition unit configured to acquire a plurality of training parameters of the vehicle and a target comfort braking force; a predicted brake generating unit configured to generate a predicted comfortable braking force from a comfortable braking model based on the plurality of training parameters; and a model parameter adjustment unit configured to adjust the plurality of model parameters of the comfort braking model based on the target comfort braking force and the predicted comfort braking force.

In some embodiments, wherein the model parameter adjustment unit comprises: a loss value determination unit configured to determine a model loss value using a loss function based on the target comfortable braking force and the predicted comfortable braking force; a gradient value determining unit configured to determine a plurality of gradient values for the plurality of model parameters using a gradient descent algorithm or the like based on the model loss value; and a model parameter second adjustment unit configured to adjust the plurality of model parameters of the comfort brake model based on the plurality of gradient values.

In some embodiments, the model update triggering unit further comprises: a model parameter deployment unit configured to deploy the plurality of model parameters of the trained comfort brake model onto the vehicle.

In some embodiments, wherein the comfort brake determination unit 804 comprises: a first model acquisition unit configured to acquire a first comfortable braking model deployed on the vehicle; and a first braking determination unit configured to determine the comfortable braking force from the first comfortable braking model based on the plurality of model parameters.

In some embodiments, wherein the first comfortable braking model is generated by a second comfortable braking model deployed at the cloud end, and the second comfortable braking model has the same model parameters as the first comfortable braking model.

In some embodiments, wherein the model parameter updating unit 802 comprises: an abnormal data determination unit configured to acquire braking behavior data, vehicle state data, and road condition data of the vehicle; and a model update prohibition unit configured to prohibit the updating of a plurality of model parameters in response to determining that at least one of the brake behavior data, the vehicle state data, and the road condition data does not satisfy a predetermined condition.

In some embodiments, wherein the plurality of vehicle parameters related to the vehicle comprises at least two of: the current speed of the vehicle; deceleration of the vehicle; the weight of the vehicle; the brake disc temperature of the vehicle; a brake disc friction coefficient of the vehicle; or a brake disc wear coefficient of the vehicle.

Fig. 9 illustrates a schematic block diagram of an example device 900 suitable for use in practicing embodiments of the disclosure. As shown, device 900 includes a processor 901 that can perform various suitable actions and processes in accordance with computer program instructions stored in a Read Only Memory (ROM) 902 loaded into a Random Access Memory (RAM) 903. In the RAM 903, various programs and data required for the operation of the device 900 can also be stored. The processor 901, the ROM 902, and the RAM 903 are connected to each other by a bus 904. An input/output (I/O) interface 905 is also connected to the bus 904.

The various methods and processes described above may be performed by the processor 901. For example, in some embodiments, the various methods and processes described above may be implemented as a computer software program tangibly embodied on a machine-readable medium. In some embodiments, some or all of the computer program may be loaded and/or installed onto device 900 via ROM 902. When the computer program is loaded into RAM 903 and executed by processor 901, one or more actions of the methods and processes described above may be performed.

The present disclosure may be methods, apparatus, systems, and/or computer program products. The computer program product may include a computer readable storage medium having computer readable program instructions embodied thereon for performing aspects of the present disclosure.

The computer readable storage medium may be a tangible device that can hold and store instructions for use by an instruction execution device. The computer readable storage medium may be, for example, but not limited to, an electronic storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of the computer-readable storage medium would include the following: random Access Memory (RAM), read Only Memory (ROM), erasable programmable read only memory (EPROM or flash memory), static Random Access Memory (SRAM), and any suitable combination of the foregoing. Computer-readable storage media, as used herein, are not to be construed as transitory signals per se, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., optical pulses through fiber optic cables), or electrical signals transmitted through wires.

The computer readable program instructions described herein may be downloaded from a computer readable storage medium to a respective computing/processing device or to an external computer or external storage device over a network, such as the internet, a local area network, a wide area network, and/or a wireless network. The network may include copper transmission cables, fiber optic transmissions, wireless transmissions, routers, firewalls, switches, gateway computers and/or edge servers. The network interface card or network interface in each computing/processing device receives computer readable program instructions from the network and forwards the computer readable program instructions for storage in a computer readable storage medium in the respective computing/processing device.

The computer program instructions for performing the operations of the present disclosure may be assembly instructions, instruction Set Architecture (ISA) instructions, machine-related instructions, microcode, firmware instructions, state setting data, or source or object code written in any combination of one or more programming languages, including an object oriented programming language such as SMALLTALK, C ++ or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The computer readable program instructions may be executed entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computer (for example, through the Internet using an Internet service provider). In some embodiments, aspects of the present disclosure are implemented by personalizing electronic circuitry, such as programmable logic circuitry, field Programmable Gate Arrays (FPGAs), or Programmable Logic Arrays (PLAs), with state information of computer readable program instructions, which can execute the computer readable program instructions.

Various aspects of the present disclosure are described herein with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the disclosure. It will be understood that each block of the flowchart illustrations and/or block diagrams, and combinations of blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer-readable program instructions.

These computer readable program instructions may be provided to a processing unit of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processing unit of the computer or other programmable data processing apparatus, create means for implementing the functions/acts specified in the flowchart and/or block diagram block or blocks. These computer readable program instructions may also be stored in a computer readable storage medium that can direct a computer, programmable data processing apparatus, and/or other devices to function in a particular manner, such that the computer readable medium having the instructions stored therein includes an article of manufacture including instructions which implement the function/act specified in the flowchart and/or block diagram block or blocks.

The computer readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other devices to cause a series of operational steps to be performed on the computer, other programmable apparatus or other devices to produce a computer implemented process such that the instructions which execute on the computer, other programmable apparatus or other devices implement the functions/acts specified in the flowchart and/or block diagram block or blocks.

The flowcharts and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of instructions, which comprises one or more executable instructions for implementing the specified logical function(s). In some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.

The foregoing description of the embodiments of the present disclosure has been presented for purposes of illustration and description, and is not intended to be exhaustive or limited to the embodiments disclosed. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the various embodiments described. The terminology used herein was chosen in order to best explain the principles of the embodiments, the practical application, or the improvement of technology in the marketplace, or to enable others of ordinary skill in the art to understand the embodiments disclosed herein.

Claims (14)

1. A method for comfortable braking of a vehicle, comprising:

updating a plurality of model parameters for the vehicle;

Determining a comfort braking force for the vehicle based on a plurality of vehicle parameters related to the vehicle and the plurality of model parameters; and

And controlling comfortable braking of the vehicle based on the comfortable braking force.

2. The method of claim 1, wherein updating a plurality of model parameters for the vehicle comprises:

determining whether to trigger an update of the plurality of model parameters based on one or more of the plurality of vehicle parameters; and

In response to determining to trigger the updating, a plurality of model parameters for the vehicle are updated.

3. The method of claim 2, wherein determining whether to trigger an update of the plurality of model parameters comprises:

Determining a target deceleration of the vehicle based on a vehicle speed of the plurality of vehicle parameters;

Determining a deceleration difference between a vehicle deceleration in the vehicle parameters and the target deceleration; and

In response to the deceleration difference being greater than a difference threshold, an update of the plurality of model parameters is triggered.

4. The method of claim 2, wherein updating the plurality of model parameters for the vehicle comprises:

Acquiring a plurality of training parameters and target comfortable braking force of the vehicle;

generating a predicted comfort braking force from a comfort braking model based on the plurality of training parameters; and

The plurality of model parameters of the comfort braking model are adjusted based on the target comfort braking force and the predicted comfort braking force.

5. The method of claim 4, wherein adjusting the plurality of model parameters of the comfort braking model comprises:

Determining a model loss value using a loss function based on the target comfort braking force and the predicted comfort braking force;

Determining a plurality of gradient values for the plurality of model parameters using a gradient descent algorithm based on the model loss value; and

Based on the plurality of gradient values, the plurality of model parameters of the comfort braking model are adjusted.

6. The method of claim 4, further comprising:

Deploying the plurality of model parameters of the adjusted comfort braking model onto the vehicle.

7. The method of claim 1, wherein determining the comfort braking force for the vehicle comprises:

Acquiring a first comfortable braking model deployed on the vehicle; and

The comfort braking force is determined by the first comfort braking model based on the plurality of model parameters.

8. The method of claim 7, wherein the first comfortable braking model is deployed based on a second comfortable braking model on a cloud server, and the second comfortable braking model has the same model architecture as the first comfortable braking model.

9. The method of claim 1, wherein updating a plurality of model parameters for the vehicle comprises:

Acquiring braking behavior data, vehicle state data and road condition data of the vehicle; and

In response to determining that at least one of the braking behavior data, the vehicle state data, and the road condition data does not satisfy a predetermined condition, disabling the updating of a plurality of model parameters.

10. The method of claim 1, wherein the plurality of vehicle parameters related to the vehicle comprises at least two of:

the current speed of the vehicle;

Deceleration of the vehicle;

The weight of the vehicle;

the brake disc temperature of the vehicle;

A brake disc friction coefficient of the vehicle; or alternatively

The brake disc wear coefficient of the vehicle.

11. An apparatus for comfortable braking of a vehicle, comprising:

A model parameter updating unit configured to update a plurality of model parameters for the vehicle;

A comfort brake determining unit configured to determine a comfort braking force for the vehicle based on a plurality of vehicle parameters related to the vehicle and the plurality of model parameters; and

A comfortable braking control unit configured to control comfortable braking of the vehicle based on the comfortable braking force.

12. A controller, comprising:

at least one processor; and

A memory coupled to the at least one processor and having instructions stored thereon that, when executed by the at least one processor, cause the controller to perform the method of any of claims 1-10.

13. A vehicle comprising the controller of claim 12.

14. A machine-readable storage medium having stored thereon machine-executable instructions, wherein the machine-executable instructions are executed by a processor to implement the method of any of claims 1 to 10.