Brake control method, model training method, and related devices

Abstract

A brake control method, a model training method and related devices. The control method comprises: obtaining first driving environment information (S701), the first driving environment information comprising information of an external environment of the vehicle; obtaining a first brake trigger instruction input by a first user (S702), the first brake trigger instruction being used to instruct brake control; and in response to the first brake trigger instruction, obtaining first brake control data based on the first driving environment information (S703), the first brake control data being used to control the brake device to brake, the first brake data corresponding to the first user. The control method generates first brake control data by obtaining first driving environment information and in response to a first brake trigger instruction, enabling the vehicle to brake smoothly and safely, and improving the brake experience of the user and the intelligent level of the vehicle.

Description

Braking control methods, model training methods and related devices Technical Field

[0001] This application relates to intelligent vehicles and artificial intelligence (AI) technology, and in particular to a braking control method, a model training method, and related devices. Background Technology

[0002] The brake pedal in a vehicle can be set to adjust its sensitivity to suit the driver's habits. However, in actual use, it's often difficult to balance brake pedal sensitivity and pedal comfort. At low sensitivity, the brake pedal offers little resistance, allowing for a large opening with a light press, and the vehicle gradually decelerates to a stop, giving the impression of delayed braking force output—the pedal feels "soft." In this case, the driver needs to apply considerable pressure to the pedal, leading to foot pain. At high sensitivity, the brake pedal offers greater resistance, requiring more force to achieve a larger opening, but the vehicle stops quickly and abruptly after a hard press—the pedal feels "hard." In this case, even a light press results in sudden deceleration, affecting passenger comfort.

[0003] Some solutions utilize the user's historical pedal pressure to adaptively adjust pedal sensitivity, making the brake pedal sensitivity match the user's driving habits. While this approach adjusts the pedal's firmness, the braking process remains mechanical, resulting in a poor user experience. For example, even with appropriate pedal sensitivity, in situations requiring emergency or rapid braking, the driver still needs to press the brake pedal forcefully. Furthermore, in cases of user misjudgment or failure to apply appropriate braking force in time, current braking systems are prone to sudden deceleration or delayed deceleration, affecting not only user comfort but also potentially leading to accidents. Summary of the Invention

[0004] This application provides a braking control method, a model training method, and related apparatus. This application can respond to a user's braking trigger instruction and determine a braking strategy for the user based on the current driving environment. The user only needs to give a braking trigger command, and the vehicle can brake smoothly and safely, improving the user's braking experience and enhancing the vehicle's intelligence level. Furthermore, the braking strategy can be personalized and associated with the user who triggers the braking, making the braking process adaptable to the user's needs and further improving the user's driving experience.

[0005] Firstly, this application provides a braking control method that can be applied to a vehicle, for example, implemented by the vehicle's control equipment (e.g., a controller), or implemented by a module within the control equipment. This module may include software modules, hardware modules, or a combination of both. For ease of description, the following explanation will use a braking control device as an example to illustrate the implementation of the braking control method.

[0006] The braking control method includes: acquiring first driving environment information, acquiring a first braking trigger indication input by a first user, and in response to the first braking trigger indication, obtaining first braking control data based on the first driving environment information.

[0007] The first driving environment information describes the environment in which the vehicle is located and may include information about the vehicle's external environment. The first braking control data indicates a braking strategy (i.e., a braking scheme). For example, the first braking control data includes one or more of the following: braking pressure data (i.e., what pressure to use for braking), braking acceleration data, etc. The first braking control is used to control the braking device. For example, the control device may control the braking device based on the first braking control data, or the control device may provide the first braking control data to the controller of the braking control device for controlling the braking device.

[0008] In this application, the braking control device can obtain a braking strategy adapted to the current environment based on the analysis of the vehicle's environment. When the user needs to brake, the user provides a braking trigger instruction (equivalent to turning on a switch), such as lightly pressing the brake pedal. The braking control device responds to the user's braking trigger instruction and automatically analyzes the current environmental information to obtain braking control data. In the above scheme, the braking control data and the force applied by the user to press the brake pedal are not linearly related. The braking control device can brake based on the current scenario. The solution provided in this application can prevent the driver from applying improper pedal force, which would result in unsafe or uncomfortable braking distance and acceleration. In other words, this application can increase driving safety and improve the driving experience. On the one hand, in scenarios requiring rapid braking, the user no longer needs to use a large amount of force to press the brake pedal, greatly reducing the possibility of foot pain caused by sudden braking, reducing driving risks, improving user comfort, and making the braking process safer and more intelligent. On the other hand, in scenarios where rapid deceleration is not required, after the user gives a braking trigger instruction, the braking control device can automatically provide a braking strategy based on the environment, making the braking process smooth and comfortable, and improving the user's driving experience.

[0009] Furthermore, the first braking control data corresponds to the first user. That is, the first braking control data is targeted at the first user. In one implementation, the process of obtaining the braking control data utilizes information about the first user, such as one or more of the first user's identity or user characteristics. These user characteristics may include one or more of age, gender, clothing, mood, or weight. For other users different from the first user, different braking control data may be output in the same scenario. In another implementation, the parameters of the algorithm or model used to obtain the braking control data are adapted to the first user, or adjusted based on the first user's identity, or adjusted based on the first user's usage habits. In other words, the algorithm or model used to obtain the braking control data can generate personalized braking control data for the first user based on driving environment information.

[0010] In the above scheme, the initial braking data is personalized and associated with the user who triggered the braking. Thus, the braking control data determined based on scenario information can adapt to the braking needs of the first user, further enhancing the user experience. For example, an aggressive braking strategy can be adopted for young and middle-aged people, while a more conservative braking strategy can be used for the elderly. Furthermore, by collecting user feedback on the executed braking operations, the parameters of the algorithm or model can be adjusted to adapt the braking control data to the current user's habits, achieving personalized braking control.

[0011] In one possible implementation of the first aspect, obtaining braking control data based on first driving environment information includes: inputting the first driving environment information into at least one model to obtain first braking control data. Further, the at least one model is trained based on the driving environment information and braking performance data corresponding to the driving environment information.

[0012] The model learns from historical data (i.e., training data), analyzes the correlations and patterns within the historical data, and performs tasks such as prediction or classification. Based on this approach, in the process of obtaining braking control based on driving environment information, since driving environment information includes multi-dimensional data and may include various unexperienced scenarios, using the model to infer scenario-adaptive braking control data ensures optimal braking acceleration and distance when dealing with complex driving environments, increasing driving safety and improving the driving experience. Furthermore, the model has strong generalization ability; using the braking control data obtained from the model can improve the accuracy of braking control decisions, thereby enhancing braking safety.

[0013] Furthermore, at least one model can be a single model, or multiple models can be used. When using multiple models, each model can undertake different tasks to jointly achieve the goal of obtaining braking control data based on driving environment information.

[0014] Furthermore, at least one model includes a machine learning model or a neural network model. For example, at least one model includes a first model, which is a large model, meaning a model with a large number of parameters and a complex structure. Exemplarily, the first model is a convolutional neural network (CNN) model, a recurrent neural network (RNN) model, or a Transformer-based model, etc.

[0015] In some cases, the model can be deployed in a separate computing unit that is adapted to the model's computational characteristics, such as having tensor computing units or matrix computing units. In this scenario, a separate computing unit can be set up in the vehicle to deploy the model, so that the model can be used to obtain braking control data, thereby making rational allocation of computing resources and improving the utilization rate of computing resources.

[0016] In another possible implementation of the first aspect, obtaining braking control data based on first driving environment information includes: obtaining first braking control data using a predefined algorithm based on the first driving environment information. For example, using a rule-based algorithm, the first braking control data is output when the first driving environment information meets a first condition. For instance, if there are no obstacles ahead and the weather is not rainy or snowy, the system first decelerates with acceleration A1 for a duration T1, and then decelerates with acceleration A2 for a duration T2, where A1 and A2 are used to exemplarily represent accelerations, and T1 and T2 are used to exemplarily represent durations. It should be understood that acceleration and deceleration are interchangeable in this document, and since acceleration has a direction, in some cases, deceleration is acceleration in the opposite direction to velocity.

[0017] In another possible implementation of the first aspect, the information about the vehicle's external environment includes one or more of the following: weather information, road surface adhesion indication information, the type of road on which the vehicle is located, emergency information about the road on which the vehicle is located, or traffic congestion information about the road on which the vehicle is located.

[0018] This includes weather information such as sunny, rainy, hail, and snowy weather. In some cases, weather information can be replaced by road surface adhesion information, or weather information can be combined with road surface adhesion indication information. Road surface adhesion information indicates the adhesion of the road surface to which the vehicle is traveling. The road category describes the location category of the road the vehicle is on. The emergency information for the road indicates whether an emergency exists, and may also indicate the specific details of the emergency if one exists. The congestion information for the road indicates whether there is congestion, or indicates the level of congestion.

[0019] In another possible implementation of the first aspect, information about the vehicle's external environment can be obtained through a network and / or through data analysis collected by the vehicle's perception system.

[0020] In another possible implementation of the first aspect, the first driving environment information further includes information about the vehicle's internal environment, which includes one or more of the following: the driver's seating posture information, the driver's seat pressure information, the characteristics of the occupants in the vehicle, the detection information of the vehicle's cabin, or the user characteristics of the driver.

[0021] The vehicle driver's seating posture information includes data collected by pressure sensors on the driver's seat and data on driver's seat posture adjustment. The driver's seat pressure data is the pressure applied by the driver to the seat, collected by pressure sensors. Occupant characteristics describe the distribution of occupants in the vehicle cabin and their personal characteristics. Cabin detection information describes the conditions within the vehicle cabin and can be obtained based on data collected by sensors in the cabin (e.g., pressure sensors, device status detection devices, cameras, or microphones). Driver user characteristics include driver age, gender, clothing, mood, attention, weight, activity level, and fatigue level.

[0022] Furthermore, the first user is the driver, that is, the first user sits in the driver's seat of the vehicle.

[0023] In yet another possible implementation of the first aspect, the vehicle further includes a brake pedal and a pedal sensor. The pedal sensor is used to detect brake pedal depressing data. Acquiring a first brake trigger indication input by the user includes: receiving a first brake trigger indication from the pedal sensor, wherein the first brake control data is independent of the brake pedal depressing force.

[0024] In the above embodiments, the first user can brake by pressing the brake pedal. The pedal sensor can detect the brake pedal pressing data and submit a brake trigger instruction. In some cases, the pressing data detected by the pedal sensor can be directly used as a brake trigger instruction, for example, when the pedal opening meets a specified condition. In other cases, the pedal sensor can generate a brake trigger instruction based on the detected data and provide it to the brake control device, for example, the pedal sensor processes the pressing data to form a brake trigger instruction. Using the user's pressing of the brake pedal to generate a brake trigger instruction conforms to the driver's braking habits, has a low user learning cost, is compatible with most existing vehicle architectures, and has good applicability.

[0025] Furthermore, the first braking control data is independent of the force applied to the brake pedal. Thus, the system can automatically generate appropriate braking control data based on driving environment information, allowing the user to apply the brake pedal with their preferred force.

[0026] In another possible implementation of the first aspect, the vehicle further includes a brake switch, such as a button or key indicating braking. A first user can operate the brake switch, which generates a brake trigger indication in response to the user's operation. The method further includes receiving a first brake trigger indication from the brake switch. In the above embodiments, the user can provide a brake trigger indication via the brake switch, for example, through a button, screen click, joystick, etc., to obtain diverse braking schemes.

[0027] In another possible implementation of the first aspect, the first braking control data includes pressure data and / or braking acceleration data. The pressure data indicates the brake pedal pressure value within a first time period, where the first time period represents a period of time; in some implementations, the length of the first time period is not fixed. For example, the pressure data can be a pressure curve, which indicates the braking force at different times within a period of time. The pressure data can characterize the deceleration strategy during braking. Furthermore, the pressure data can be matched with pedal pressure, and since the braking control of most vehicles is based on pedal pressure, the above solution is compatible with existing braking control processes and has high applicability.

[0028] Braking acceleration data is used to indicate braking acceleration within a first time period, which also indicates a period of time. For example, braking acceleration data can be a braking acceleration curve, which indicates braking acceleration at different times within a period of time.

[0029] Braking acceleration data can also characterize deceleration strategies during braking. Using braking deceleration as braking control data makes it easier to control braking distance and braking comfort, thus improving the safety and comfort of the braking process.

[0030] In another possible implementation of the first aspect, the method further includes: controlling the braking device to brake based on the first braking control data.

[0031] For example, the braking control device has the capability to control the braking device and can control the braking device to perform braking based on braking control data. The braking control device outputs a control signal to the braking device, which is generated based on the first braking control data.

[0032] As another example, the braking control device is an external control device of the braking system, connected to the controller inside the braking system. In this case, the braking control device outputs first braking control data to the braking system, and the controller inside the braking system generates a control signal based on the first braking control data to control the actuators in the braking system to perform braking.

[0033] In another possible implementation of the first aspect, the method further includes: controlling the braking device to brake based on first braking control data, and acquiring first perception data of a first user of the vehicle during a second time period after braking begins. Further, the braking control device updates first model parameters of at least one model, the first model parameters being associated with the first user, based on the first perception data.

[0034] In the above implementation, the braking control device senses user feedback on braking control data and updates the model parameters based on this feedback. This allows for personalized adjustments to the model parameters, continuously improving the model and ensuring that its subsequent outputs better align with user habits. The improved model incorporates user habits and preferences, better adapting to user braking patterns and enhancing the user experience.

[0035] In some possible implementations, the perceived data includes one or more dimensions such as the pressure applied by the user to the seat, the user's facial expressions, the user's voice, and the user's emotions. Using multi-dimensional data allows the braking control device to more accurately analyze the user's feedback to the automatically generated braking strategy, improving the model's relevance to the user and enhancing the user experience.

[0036] In another possible implementation of the first aspect, the perceived data includes pressure data collected by at least one pressure sensor located at the driver's seat of the vehicle. Based on the first perceived data, updating the first model parameters of at least one model includes: adjusting the first model parameters of at least one model in the convergence direction when the fluctuation of the pressure data meets a preset fluctuation condition; and / or, adjusting the first model parameters of at least one model in the opposite direction of the convergence direction when the fluctuation of the pressure data does not meet a preset fluctuation condition.

[0037] In the above implementation, to incorporate the driver's personalized braking preferences, the braking control device can perform adaptive training (or fine-tuning) on ​​the model. Taking the pressure information from the seat pressure sensor as the feedback signal during model adjustment as an example, during adaptive training, the braking control data output by the model is used to control the braking device. During braking, if the parameters of the seat pressure sensor are stable (i.e., the pressure data meets the fluctuation condition), it indicates that the driver is adapting to the braking strategy output by the current model, and the model parameters are adjusted in the direction of fitting convergence. Conversely, if the seat pressure sensor parameters change significantly (e.g., the parameters of sensors in the leg rest and lumbar support fluctuate greatly), it indicates that the driver is not adapting to the braking strategy output by the current model and is attempting to press the brake pedal to change the braking strategy; in this case, the model parameters are adjusted in the opposite direction of fitting convergence.

[0038] In some cases, the process of outputting braking control data and adjusting model parameters based on user perception data can be executed multiple times. In this way, the model can undergo multiple iterations to obtain an optimal model that incorporates user preferences.

[0039] Furthermore, the model parameters after adaptive training are related to the driver account logged in during adaptive training or to the identified driver's seat user. The model parameters after adaptive training can be associated with the driver account or the user, for example, stored in the corresponding driver account. Taking the currently logged-in driver account as the first user's account as an example, the model parameters after adaptive training are associated with the first user.

[0040] In another possible implementation of the first aspect, the method further includes: the braking control device acquiring third driving environment information, the third driving environment information including information about the external environment of the vehicle; the braking control device acquiring a second braking trigger indication input by a first user, the second braking trigger indication indicating braking control; and, in response to the second braking trigger indication, obtaining third braking control data based on the third driving environment information using at least one model, wherein the at least one model is adaptively trained, and the model parameters applied by the at least one model are model parameters corresponding to the first user.

[0041] In the above embodiments, after adaptive training, the braking control device can use the adaptively trained model to obtain the optimal braking control strategy based on the current environmental information, so that the braking control strategy meets the user's preferences.

[0042] In another possible implementation of the first aspect, after controlling the braking device based on the first braking control data, the method further includes: detecting the vehicle's operating status; if the vehicle is operating normally, then performing the step of "updating the first model parameters of at least one model based on the first perception data". Alternatively, if the vehicle is operating abnormally, then the step of "updating the first model parameters of at least one model based on the first perception data" is not performed. Abnormal operation includes situations such as a vehicle collision or a vehicle malfunction, such as a faulty braking device. Based on the above implementation, if an abnormality occurs after controlling the braking device using automatically generated braking control data, it indicates that the model parameters may be abnormal; in this case, the step of adjusting the model parameters based on user feedback is not performed.

[0043] In another possible implementation of the first aspect, before inputting the first driving environment information into at least one model to obtain the first braking control data, the method further includes: obtaining the identification information of a first user of the vehicle, and determining, based on the identification information of the first user of the vehicle, first model parameters applied by at least one model, the first model parameters being associated with the first user.

[0044] Identity verification information includes one or more of the following: biometric information (such as face, iris, voiceprint, or fingerprint), account information, etc. In the above method, the braking control device sets the model parameters applied by at least one model to the first model parameters corresponding to the first user, based on the first user's identity verification information. Thus, the braking control data output by at least one model can match the first user's braking preferences, improving the first user's braking experience.

[0045] In another possible implementation of the first aspect, after acquiring the first brake trigger indication input by the user, the method further includes: detecting whether the first brake trigger indication is a false trigger. Based on the above implementation, after receiving the brake trigger indication input by the user, the brake control device can perform false trigger detection on the current brake trigger indication, so that the brake control is executed when the user actually needs to brake, thereby improving driving safety and comfort.

[0046] In another possible implementation of the first aspect, obtaining the first braking control data based on the first driving environment information includes: obtaining the first braking control data based on the first driving environment information when the first braking trigger indication is not a false trigger. In the above implementation, when the first braking trigger indication is not a false trigger, braking control data is generated, so that the braking operation is performed when braking is actually required, thereby improving driving safety and comfort.

[0047] In another possible implementation of the first aspect, detecting whether the first brake trigger indication is falsely triggered includes: determining whether the first brake trigger indication is falsely triggered based on brake pedal depress data and the driver's depressing habits data.

[0048] In the above embodiments, the brake control device can analyze the brake pedal depress data and the driver's depressing habits to determine whether the first brake trigger indication is falsely triggered. For example, if the pressure change curve of the user's current depress data differs significantly from the pressure curve the user commonly uses, then the first brake trigger indication is determined to be falsely triggered.

[0049] In another possible implementation of the first aspect, detecting whether the first brake trigger indication is falsely triggered includes: determining whether the first brake trigger indication is falsely triggered based on brake pedal depress data and vehicle driving scenario information.

[0050] The vehicle's driving scenario information includes one or more of the following: obstacle information, map information or road conditions, congestion level, and emergency status. Analyzing this driving scenario information can more accurately determine whether the brake trigger indication is false, thus helping to improve vehicle driving safety.

[0051] In another possible implementation of the first aspect, detecting whether the first brake trigger indication is falsely triggered includes: determining whether the first brake trigger indication is falsely triggered based on brake pedal depress data, driver's depressing habits data, and vehicle driving scenario information.

[0052] In the above implementation, by combining vehicle driving scenario information and driver's pedaling habits, it is possible to more accurately determine whether the brake trigger indication is falsely triggered, which helps to improve vehicle driving safety.

[0053] In another possible implementation of the first aspect, detecting whether the first brake trigger indication is falsely triggered includes: determining whether the first brake trigger indication is falsely triggered based on the number of brake trigger indications received within a second time period, wherein the brake trigger indications received within the second time period include the first brake trigger indication.

[0054] When a user needs to brake, the driver typically triggers the brakes continuously or multiple times. In the above implementation, the number of times the driver triggers the brakes within a continuous time period is monitored to determine whether the driver has mistried the braking. For example, if the driver gives two consecutive brake trigger instructions within a limited time (e.g., within 2 seconds, within 1 second, etc.), it is determined that the driver intends to brake, i.e., the brake trigger instruction is not mistried. Optionally, the second time period can be predefined, or the length of the second time period is related to the vehicle's environment. For example, when the vehicle is at high speed, braking scenarios are less likely to occur, so the second time period can be set shorter. Conversely, when the vehicle is in urban areas, braking is more frequent, so the second time period can be set longer, making it less likely that the brake triggering will be detected as mistried.

[0055] In another possible implementation of the first aspect, the vehicle further includes a driver monitoring system that detects whether the first brake trigger indication is falsely triggered, including: acquiring second perception data of the driver of the vehicle after the generation time of the first brake trigger indication, and determining whether the first brake trigger indication is falsely triggered based on the second perception data.

[0056] Understandably, when a driver accidentally triggers the brakes, the driver may be in a state of tension, panic, excitement, or fear. In the above implementation, combining the driver's perceptual information with the determination of accidental brake triggering can improve the accuracy of accidental brake triggering detection.

[0057] For example, the second perception data includes one or more of the following: driver's gaze recognition data, driver's fatigue level detection data, driver's attention detection data, driver's facial expression analysis data, driver's image, driver's action analysis, or driver's voice data.

[0058] In another possible implementation of the first aspect, the above-mentioned "obtaining first braking control data based on first driving environment information in response to a first braking trigger indication" is performed when the first function is activated.

[0059] As one possible design, the braking control device, in response to a first braking trigger indication, obtains first braking control data based on first driving environment information when the first function is activated. It should be understood that the name of this first function can be designed based on actual conditions, such as adaptive braking function, automatic braking function, scenario braking, etc.

[0060] As another possible design, before obtaining the first braking control data based on the first driving environment information, the method further includes: receiving a first operation instruction input by a user, the first operation instruction being used to instruct the activation of a first function.

[0061] As another possible design, the method also includes: outputting a first prompt message, which indicates that the first function is active. The first prompt message can be output in the form of interface, voice, light, vibration, etc.

[0062] In another possible implementation of the first aspect, the method further includes: deactivating the first function, i.e., exiting the first function, when an abnormal vehicle operating state is detected. Alternatively, when an abnormal vehicle operating state is detected, the operation of "obtaining braking control data based on driving environment information" is no longer performed. For example, in a scenario where braking control data is obtained using a model, at least one of the aforementioned models is stopped from being enabled when an abnormal vehicle operating state is detected. Further, all or some of the at least one model may be stopped from being enabled.

[0063] In some solutions, the vehicle's operating status is obtained by analyzing its driving data. As one possible design, the braking control device acquires the vehicle's driving data and, based on this data, disables at least one model. That is, the braking control device can analyze the vehicle's driving data and disable models based on this data, thereby making model application safer, improving vehicle safety, and enhancing system stability.

[0064] Secondly, this application provides a model training method applied to a device with computing capabilities, such as a computing device or a module within a computing device. The following description uses a model training device as the executing entity.

[0065] The method includes: acquiring multiple training data sets; training at least one model based on the multiple training data sets; using the at least one model to obtain target braking control data based on current driving environment information in response to a braking trigger indication; and using the target braking control data to control the vehicle's braking device. Each training data set includes driving environment information, which includes information about the vehicle's external environment.

[0066] In the above scheme, the model training device can train at least one model based on the vehicle's driving environment information and its braking performance in these environments. This allows the at least one model to adaptively output optimal braking control data based on the current environment, thereby achieving optimal braking performance. When the model is applied, the user provides a braking trigger instruction (equivalent to turning on a switch), and at least one model can perform braking control based on the current driving scenario information. This model can prevent the driver from applying inappropriate pedal force, which could lead to unsafe or inappropriate braking distance and acceleration.

[0067] Because driving environment information includes multi-dimensional data and may include various unfamiliar scenarios, using models to infer scenario-appropriate braking control data can ensure optimal braking acceleration and distance when dealing with complex driving environments, increasing driving safety and improving the driving experience. Furthermore, the model has strong generalization capabilities; using the braking control data obtained from the model can improve the accuracy of braking control decisions, thereby enhancing braking safety.

[0068] In one possible implementation of the second aspect, at least one model is trained based on multiple training data, including: inputting second driving environment information into at least one model to be trained to obtain second braking control data, acquiring second braking performance data when controlling vehicle braking with the second braking control data under the second driving environment information, and adjusting the first parameter of the at least one model to be trained based on the second braking performance data.

[0069] In another possible implementation of the second aspect, adjusting the first parameter of at least one model to be trained based on the second braking performance data includes: detecting the vehicle's operating state when the vehicle is controlled to brake with the second braking control data, and when the vehicle's operating state is detected to be normal, adjusting the first parameter of at least one model to be trained based on the second braking performance data.

[0070] In another possible implementation of the second aspect, braking performance data includes braking acceleration and / or braking distance.

[0071] In another possible implementation of the second aspect, the information about the vehicle's external environment includes one or more of the following: weather information, the type of road on which the vehicle is located, emergency information about the road on which the vehicle is located, or traffic congestion information about the road on which the vehicle is located.

[0072] In another possible implementation of the second aspect, the first driving environment information further includes information about the vehicle's internal environment, which includes one or more of the following: the driver's seating posture information or the seat pressure information of the driver's seat.

[0073] In another possible implementation of the second aspect, the first braking control data includes pressure data and / or braking acceleration data. The pressure data is used to indicate the brake pedal pressure value during a first time period, and the braking acceleration data is used to indicate the braking acceleration during the first time period.

[0074] Thirdly, this application provides a braking control device, the computing device including an acquisition unit and a processing unit. The acquisition unit is used to acquire data, for example, to perform one or more operations such as acquisition, reception, or collection. The processing unit is used to process data and perform operations, for example, to implement one or more operations such as data processing, model running, execution, and control. This braking control device is used to implement the method described in the first aspect or any possible embodiment of the first aspect.

[0075] Fourthly, this application provides a model training apparatus, which includes an acquisition unit and a training unit. The acquisition unit is used to acquire data, for example, to perform one or more operations such as acquisition, reception, or collection. The training unit is used to process data and perform operations, for example, to implement one or more of the aforementioned operations such as data processing, model running, training, execution, and control. This model training apparatus is used to implement the methods described in the second aspect or any possible implementation of the second aspect.

[0076] Fifthly, this application provides a control device, including a memory and at least one processor. The memory stores computer instructions, and the at least one processor invokes the computer instructions to implement the method described in the first aspect or any possible embodiment of the first aspect.

[0077] Sixthly, this application provides a computing device including a memory and a processor. The memory stores computer instructions, and the processor invokes the computer instructions to implement the method described in the second aspect or any possible embodiment of the second aspect.

[0078] In a seventh aspect, this application provides a braking system, including a braking device, and further including a braking control device of the third aspect or a control device of the fifth aspect, wherein the braking device is used to perform braking based on braking control data provided by the braking control device or the control device.

[0079] Eighthly, this application provides a vehicle that includes the control equipment of the fifth aspect or the braking system of the seventh aspect.

[0080] Ninthly, this application provides a computer-readable storage medium storing program instructions that, when executed by at least one processor, implement the method described in the first aspect or any possible implementation of the first aspect, or implement the method described in the second aspect or any possible implementation of the second aspect.

[0081] In a tenth aspect, this application provides a computer program product, including program instructions or executable computer program code, which, when executed by at least one processor, implements the method described in the first aspect or any possible implementation of the first aspect, or implements the method described in the second aspect or any possible implementation of the second aspect.

[0082] The beneficial effects of the solutions in the second to tenth aspects of this application can be found in the beneficial effects of the solutions described in the first aspect. Attached Figure Description

[0083] The accompanying drawings used in the description of the embodiments will be briefly introduced below.

[0084] Figure 1 is a schematic diagram of a vehicle's system architecture;

[0085] Figure 2 is a schematic diagram of a driver controlling a vehicle;

[0086] Figure 3 is a schematic diagram of pedal opening, pressure curve and braking acceleration curve;

[0087] Figure 4 is a flowchart illustrating a model training method provided in an embodiment of this application;

[0088] Figure 5 is a schematic diagram of the model training process provided in an embodiment of this application;

[0089] Figure 6 is a schematic diagram of a neural network model provided in an embodiment of this application;

[0090] Figure 7 is a schematic flowchart of a braking control method provided in an embodiment of this application;

[0091] Figure 8 is a schematic diagram of a model-based braking control process provided in an embodiment of this application;

[0092] Figure 9 is a schematic diagram of a braking control device provided in an embodiment of this application;

[0093] Figure 10 is a schematic diagram of a model training device provided in an embodiment of this application;

[0094] Figure 11 is a schematic diagram of the structure of a control device provided in an embodiment of this application;

[0095] Figure 12 is a schematic diagram of the structure of a computing device provided in an embodiment of this application. Detailed Implementation

[0096] The following is a brief introduction to the terminology that may be used in this application:

[0097] A model, also known as an artificial intelligence (AI) model or machine learning model, is a set of functions and parameters learned after training on training data. It is used to perform specific tasks, such as prediction, classification, or other data processing tasks.

[0098] Model training is inseparable from machine learning technology. Machine learning is the science of training computer programs or systems to perform tasks without explicit instructions. Classified by learning method, machine learning can be divided into several categories, including supervised learning, unsupervised learning, and self-learning. Deep learning is a subset of machine learning, based on deep neural network models and methods.

[0099] The following section introduces one type of vehicle to which this application can be applied and its usage scenario.

[0100] Please refer to Figures 1 and 2. Figure 1 is a schematic diagram of a vehicle system architecture, and Figure 2 is a schematic diagram of a driver controlling the vehicle. As shown in Figures 1 and 2, the vehicle 10 may include a power system 11 and a braking device 12, and optionally also include a sensor system 13, control equipment 14, or peripheral equipment 15, etc. Wherein:

[0101] The power system 11 provides power to the vehicle 10, and may include one or more of the following: an engine, a power battery, etc.

[0102] Braking device 12 can refer to a device for slowing down the speed of vehicle 10, and may also be called a braking device. It may include a speed reducer or other structural components used for vehicle deceleration. In some embodiments, braking device 12 may use friction to slow down the movement of the wheels, thereby reducing the speed of the vehicle.

[0103] Some vehicles also include a brake pedal (not shown), which is configured to be movable under force. For example, the driver can press or release the brake pedal, using the braking device 12 to control the speed of the vehicle 10. Furthermore, the vehicle also includes a pedal sensor (not shown), which detects brake pedal depressing data, such as pedal opening, pedal depressing speed, pedal depressing force, or one or more other data points. In some cases, the pedal sensor may be part of a sensor system 13.

[0104] The sensor system 13 may include several detection devices (including probes) that measure information and convert the measured information into electrical signals or other desired forms of information output according to a certain rule. As shown in Figure 1, the sensor system 13 of the vehicle 10 includes one or more of the following detection devices: image sensor 131, voice system 132, lidar 133, radar 134, pressure sensor 135, temperature sensor 136, or positioning system 137, etc., optionally including the aforementioned pedal sensor. Some of these detection devices are described below by way of example:

[0105] Image sensor 131 is used to capture images, including pictures and videos. In some specific implementations, image sensor 131 includes, but is not limited to, dashcams, cameras, or other components used for taking pictures / photographs. Optionally, image sensor 131 can be configured to capture images of the exterior of the vehicle to obtain information about the vehicle's surrounding environment. Alternatively, image sensor 131 can be configured to capture images of the vehicle's interior, such as images of the driver and the cabin. For example, a driver monitoring system (DMS) is deployed in the vehicle, and the DMS system includes image sensor 131, as shown in Figure 2. This image sensor 131 can be positioned facing the driver, and when activated, it can continuously and in real time acquire images in the driver's direction. As another example, a cockpit monitoring system (CMS) is deployed in the vehicle, which can acquire images of the cabin. Of course, in specific implementations, the vehicle also includes multiple image sensors 131 to simultaneously capture images of the vehicle's interior and exterior.

[0106] The voice system 132 is used to collect sound information. For example, the voice system may include or be connected to a microphone 153. In some embodiments, the voice system 132 also includes a speaker 152 for emitting sound. Furthermore, the voice system can interact with a user, such as receiving user input (e.g., collecting cockpit voice), and / or providing voice prompts to the user, thereby engaging in voice interaction with the user. In some embodiments, the voice system 132 can be used to collect cockpit voice.

[0107] LiDAR 133 and radar 134 are devices that detect objects using electromagnetic waves (including light). They can obtain relevant information about targets in the object space by emitting signals and receiving echoes, including one or more of the target's distance (or depth), angle, speed, reflectivity, and color. For example, referring to Figure 2, LiDAR 133 can be configured to face outwards from the vehicle to detect targets around the vehicle. In some embodiments, LiDAR 133 and radar 134 can be used to detect information about the vehicle's surrounding environment, such as static environmental information and dynamic environmental information.

[0108] Pressure sensor 135 is used to detect applied pressure and can be installed in the vehicle seat. Referring to Figure 2, pressure sensor 135 may be located in the seat back, seat panel, leg rest, headrest, armrest, etc. In some solutions, the measurement data from the pressure sensor in the vehicle seat can be used to analyze the occupant's posture and sense occupant feedback. Temperature sensor 136 is used to measure temperature. Positioning system 137 is a device for acquiring location information, which can be used to achieve real-time vehicle positioning and provide the vehicle's geographical location information. Positioning systems include, for example, the Global Positioning System (GPS) or the BeiDou Navigation Satellite System.

[0109] Peripheral device 15 may include several components, such as the human-machine interaction (HMI) 151, speaker 152, microphone 153, etc., as shown in the figure. An HMI is a device that connects to input and / or output devices to enable human-machine information interaction, including but not limited to displays (such as vehicle central control screens, streaming rearview mirrors, instrument panels, head-up displays (HUDs), light field screens, or projectors, etc.) and touchscreens. In some solutions, speakers and microphones can also be considered HMIs. Speaker 152, also called a loudspeaker, is used to convert audio electrical signals into sound signals. The vehicle listens to music or hands-free calls through speaker 152. Microphone 153, also called a microphone or transducer, is used to convert sound signals into electrical signals. When making a phone call or sending a voice message, the user speaks near microphone 153, and microphone 153 inputs the sound signal into the microphone.

[0110] Control device 14 is a computing device that may include one or more processors. The processors can run programs or corresponding instructions to achieve specific functions (described below). Exemplarily, the computing device may be a controller, a mobile data center (MDC) (or autonomous driving domain controller), a domain controller (DC), an electronic control unit (ECU), etc., where the DC may include a motion domain controller (MDC), a vehicle domain controller (VDC), etc. In some solutions, control device 14 may not be located inside the vehicle, for example, it may be located in the cloud, on a roadside device, or in a data center.

[0111] As one possible implementation, the control device 14 can be combined with one or more other components in the vehicle, such as the power system 11, braking device 12, and sensor system 13 in the sensor system 13, to achieve driver assistance functions. For example, the control device 14 can control the braking device 12 of the vehicle 10 to perform braking based on data collected by the sensor system 13, that is, it acts as a controller for the braking system. Alternatively, the control device 14 can be connected to a controller in the braking system.

[0112] In some designs, the vehicle also includes a memory to provide storage space. For example, the memory may include volatile memory, such as RAM. Alternatively, the memory may include non-volatile memory, such as read-only memory (ROM), flash memory, hard disk drive (HDD), or solid-state drive (SSD). Combinations of these types of memory are also possible. Optionally, the memory may also store information such as road maps, driving routes, and sensor data.

[0113] It should be noted that Figure 1 above is only a schematic diagram of one possible functional framework of vehicle 10. In practical applications, vehicle 10 may include more or fewer systems or components, and this invention does not limit this. For example, vehicle 10 may also include a power supply or a communication system, etc.

[0114] Referring to Figure 2, when driving a vehicle, the user controls the vehicle's braking by using the brake pedal. The braking strategy is related to the user's action of pressing the brake pedal. See part (a) of Figure 3; at time T0, the user has not pressed the brake pedal, and the brake pedal opening is not 0. After the user presses the pedal, the pedal opening will change. See parts (b) and (c) of Figure 3; at time T2, the pedal is pressed down by an angle β1, and at time T2 it is pressed down by an angle β2. The controller in the braking system can obtain a pressure curve or braking acceleration curve based on the pedal opening data to control the vehicle's braking. The pressure curve is shown in part (d) of Figure 3, and the braking acceleration curve is shown in part (e) of Figure 3. The pressure curve or braking acceleration curve reflects the vehicle's braking strategy. Since both the pressure curve and the braking acceleration curve are related to the change in brake pedal opening, when the user needs to decelerate quickly, they often need to press the brake pedal with greater force, which can easily cause foot pain and low comfort.

[0115] To improve braking comfort, brake pedal sensitivity can be adjusted by the user to suit their driving habits. However, in reality, brake pedal sensitivity and passenger comfort are often difficult to balance. Since braking relies on the driver's manual application of pressure and pedal depth, even with good brake pedal sensitivity, different braking strategies are required in different environments, making it difficult for the vehicle's braking response to meet user expectations. For example, if the pedal sensitivity is set too high, a light touch will result in sudden deceleration, significantly reducing passenger comfort when rapid braking is not necessary. Conversely, if the pedal sensitivity is set too low, the user will need to apply greater force to the pedal in situations requiring rapid deceleration, leading to foot pain.

[0116] In some solutions, to adapt braking sensitivity to the user's driving habits, the controller adaptively adjusts the brake pedal sensitivity based on the user's braking habit data. The controller can collect the force and depth of the user's brake pedal operation, as well as the corresponding vehicle acceleration and speed, and analyze this data to obtain the optimal pedal sensitivity. However, this solution can only adjust the braking sensitivity to suit the user. In actual braking, because different scenarios require different braking strategies, braking remains mechanical, resulting in a poor user experience. For example, in scenarios requiring emergency braking or rapid braking, the driver still needs to press the brake pedal forcefully. Furthermore, in cases where the user misjudges or fails to apply appropriate braking force in time, the current braking system is prone to sudden deceleration or delayed deceleration, which not only affects user comfort but may also lead to accidents.

[0117] In view of this, this application provides a braking control method, a model training method, and related apparatus, capable of responding to user-provided braking trigger instructions and obtaining braking control data based on the current driving environment. When the user needs to brake, the user provides a braking trigger instruction (equivalent to turning on a switch), such as lightly pressing the brake pedal. The braking control device responds to the user's braking trigger instruction, automatically analyzes the current environmental information, and obtains braking control data. This prevents the driver from applying inappropriate pedal force, which could lead to unsafe or uncomfortable braking distances and accelerations, enabling the vehicle to brake smoothly and safely, increasing driving safety and improving the driving experience. Furthermore, the braking strategy is personalized and associated with the user who triggers the braking, allowing the braking process to adapt to the user's needs, further enhancing the user's driving experience.

[0118] When the braking indication is given through the brake pedal, the driver only needs to lightly press the brake pedal with a comfortable (or habitual) force to apply light pressure. Based on the input of driving environment information, this application can intelligently obtain braking control data to achieve the optimal braking distance and / or acceleration.

[0119] In some solutions, driving environment information is input into the model, which processes the information and outputs braking control data. This model is trained based on the driving environment information and the corresponding braking performance data.

[0120] To facilitate understanding, the model training process will be introduced first. Model training can be implemented using computing devices, which are devices with computing capabilities. Furthermore, computing devices concentrate a significant amount of computing resources, which can include physical devices or virtual units or software modules. For example, a computing device may include one or more physical servers, such as blade servers or rack servers. Alternatively, a computing device may contain one or more computing instances, which are virtualizations of computing resources, such as virtual machines or containers. In some solutions, the computing device can be a cloud service that provides corresponding interfaces. Devices or services outside the cloud service can interact with it by calling these interfaces, and the cloud service can also transmit information to external devices or services through these interfaces.

[0121] Please refer to Figure 4, which is a flowchart illustrating a model training method provided in an embodiment of this application. The following description uses a model training device as the executing entity, which has computational capabilities. As shown in Figure 4, the model training method may include steps S401 and S402. It should be understood that, for ease of description, the method is described in the order of S401 to S402, but this embodiment does not limit the order of execution, execution time, or number of executions of the above one or more steps. S401 to S402 are detailed below:

[0122] S401, the model training device acquires multiple training data.

[0123] Each training data point includes driving environment information, which may include one or more of the following: information about the vehicle's external environment and internal environment. Examples are provided below:

[0124] Information 1: Information about the vehicle's external environment includes one or more of the following: weather information, road surface traction indicator information, the type of road the vehicle is on, emergency information for the road the vehicle is on, or traffic congestion information for the road the vehicle is on. This information describes the vehicle's external environment from different dimensions.

[0125] This includes weather information such as one or more weather categories, such as sunny, rainy, hail, or snowy. Weather information reflects road surface traction; for example, wet roads are slippery in rainy weather, making hydroplaning more likely, while dry roads are slippery in sunny weather, and roads are more slippery in hail or snow. In some cases, weather information can be replaced by road surface traction information, or it can include road surface traction indicator information, in which case the indicator is provided by a separate parameter. Furthermore, weather information may also include information on air pollutants, etc.

[0126] Road surface adhesion information is used to indicate the adhesion of a vehicle over a road surface. For example, road surface adhesion information may include one or more of the following: road surface smoothness information, road surface type (e.g., cement road, paved road, gravel road, or dirt road), water accumulation information on the road surface, and elevation information of sampling points on the road surface.

[0127] The road category describes the location of the road where the vehicle is situated. For example, the road category may include one or more of the following: traffic lights, uphill / downhill, intersection, underpass, or mountain road.

[0128] Emergency information on the road where the vehicle is located indicates the presence of an emergency, and if an emergency exists, it specifies the details of the emergency. For example, emergency information may include one or more of the following: a sudden appearance of an obstacle (i.e., obstacle information), road construction, illegal parking, running a red light, or an accident ahead. Further, obstacle information may include one or more of the following: obstacle location, obstacle size, obstacle type, etc.

[0129] Traffic congestion information for the road where the vehicle is located is used to indicate whether there is congestion or to indicate the degree of congestion (e.g., congestion level). For example, congestion information may include one or more of the following: no congestion, moderate congestion, or severe congestion.

[0130] Information 2, the information about the vehicle's internal environment includes one or more of the following: the driver's seating posture information, the driver's seat pressure information, the characteristics of the occupants, the detection information of the vehicle's cabin, the user characteristics of the driver, or the vehicle's own motion information, etc. This information describes the vehicle's internal environment from different dimensions.

[0131] The driver's seating posture information includes data collected by pressure sensors in the driver's seat and data on the driver's seat posture adjustment. The pressure sensors can be located in one or more positions, such as the driver's seat back, headrest, seat base, or leg rest. The driver's seat posture adjustment data includes one or more of the following: backrest angle, seat fore-aft angle, seat base angle, or leg rest angle.

[0132] The driver's seat pressure data is the pressure exerted by the driver on the seat, collected by pressure sensors. These pressure sensors can be located in one or more positions, such as the driver's seat back, headrest, seat base, and leg rest. It should be understood that there may be some overlap between the driver's seat pressure information and the driver's posture information; however, the driver's seat pressure information is primarily used to analyze the pressure exerted on the seat by the driver, while the driver's posture information is primarily used to indicate the driver's seating position.

[0133] Occupant characteristics in a vehicle describe the distribution of occupants in the vehicle's cabin and their personal characteristics. For example, occupant characteristics include whether there are occupants other than the driver, and their positions (e.g., in the driver's seat, in the back seat, etc.). Other occupant characteristics include user characteristics such as whether the occupants are children, elderly, their age, gender, and their state (e.g., sleeping, tired, watching a movie, talking, or making a phone call, etc.).

[0134] Vehicle cabin detection information describes the situation within the vehicle cabin and can be obtained based on data collected by sensors in the cabin (such as pressure sensors, equipment status detection devices, cameras, microphones, etc.). For example, vehicle cabin detection information includes one or more of the following: pressure information of the seats in the cabin (or information indicating the presence of occupants or items on the seats), whether the occupants are wearing seat belts, and the usage status of the projectors in the cabin.

[0135] The user characteristics of a vehicle's driver include one or more of the following: driver's age, gender, clothing, mood, attention, weight, activity level, and fatigue level.

[0136] The vehicle's own motion information includes one or more of the following: vehicle speed, acceleration, current travel duration, and current travel distance.

[0137] It should be understood that the training data mentioned above, when actually collected, is collected with the user's consent and within the scope of laws and regulations. Furthermore, the training data is stored, transmitted, and used in compliance with laws and regulations.

[0138] In some possible implementations, each piece of training data includes information from at least two dimensions of the above information, with each piece of information corresponding to one or more feature dimensions. Furthermore, the driving environment information in the training data includes information from at least three dimensions of the above information. The more dimensions of the driving environment information, the more comprehensive the description of the current scenario. Using multi-dimensional information to train the model enables the model to comprehensively analyze the current driving environment information, infer appropriate braking strategies for the current driving environment, and improve the accuracy and comfort of the braking strategy.

[0139] Optionally, the dimensions of information included in different training data can be different. For example, training data Tra1 includes driver's riding posture information, driver's seat pressure data, weather information, road type, road emergency information, and congestion information, while training data Tra2 includes weather information, road type, road emergency information, and congestion information.

[0140] In some possible implementations, the training data can be collected during vehicle operation. For example, taking vehicle V1 as an example, vehicle V1 can collect the following data:

[0141] Data 1: Pressure applied by the brake pedal to the driver. For example, a sensor integrated into the brake pedal (i.e., a brake pedal sensor) can collect the pressure applied by the driver to the brake pedal as brake pedal pressure data. The data collected by the brake pedal sensor can be acquired, transported, or stored by computing or control devices in the vehicle.

[0142] Data 2, based on pressure sensors installed on the seat and information about the seat's tilt angle, determines the driver's seating posture. This data can be used as a posture factor, indicating the characteristic dimension of the driver's seating posture.

[0143] Data 3, based on data collected by pressure sensors installed on the seat, determines the driver's seat pressure data. This data can be used as a seat pressure factor, which indicates the characteristic dimension of driver's seat pressure.

[0144] Data 4, based on weather information obtained from solar and rain sensors and weather services, can be used as a weather factor, that is, a feature dimension indicating the weather.

[0145] Data 5, obtained through data collected by a perception system (e.g., identifying traffic lights in the environment) and map applications (e.g., Huaban Map), identifies the type of road the vehicle is on, including information such as traffic lights, inclines / declines, intersections, culverts, and mountain roads. This data can serve as a road category factor, indicating the feature dimension of the road category.

[0146] Data 6, obtained through data collected by a sensing system and map applications (such as Huaban Map), provides information on traffic congestion on the roads where vehicles are located. This information may be categorized as no congestion, moderate congestion, or severe congestion. This data can serve as a road congestion factor, indicating the degree of congestion.

[0147] Data 7, obtained through data collected by a perception system (e.g., obstacle information) and map applications (e.g., Huaban Map), provides information on emergency situations on the road where the vehicle is located. Examples of emergency information include sudden obstacles, road construction, illegal parking, running red lights, and accidents involving vehicles ahead. This data can serve as an emergency factor, indicating the characteristic dimension of an emergency situation.

[0148] For data 1 to data 7 collected on vehicle V1, the pedal pressure data is used to determine the data related to the braking process among the data collected by vehicle V1. Data 2 to data 7 can be used as training data for the model. In some solutions, vehicle V1 can upload the aforementioned data 2-7, and optionally also data 1, to the cloud. The cloud can store and clean the above data, and optionally, perform manual processing to obtain training data for model training.

[0149] In some possible implementations, the vehicle also collects braking performance data under various driving conditions, including one or more of parameters such as braking distance and braking acceleration, to describe the braking process. Alternatively, the braking performance data can be constructed by software or manually. The aforementioned braking performance data can serve as labels during model training.

[0150] S402, the model training device trains at least one model based on multiple training data, and the at least one model is used to obtain braking control data based on driving environment information.

[0151] Please refer to Figure 5. Each training data item in the multiple training datasets may include driving environment information, which may include information from one or more feature dimensions. This training data is input into the model to be trained, with braking performance data serving as the label for model training. Braking performance data includes one or more of the following: braking distance and braking acceleration.

[0152] In one possible implementation, the model to be trained (for ease of distinction, referred to as at least one model) includes a machine learning model, such as a supervised learning-based machine learning model, with the brake performance data serving as labels for the supervised learning.

[0153] Taking one training iteration as an example, the model training device inputs a set of driving environment information (referred to as "second driving environment information" for easy distinction) into at least one model to be trained, obtaining corresponding braking control data (referred to as "second braking control data" for easy distinction). That is, at least one model analyzes and outputs second braking control data based on the second driving environment information. The model training device acquires braking performance data (referred to as "second braking performance data") when controlling the vehicle's braking under the second driving environment information using the second braking control data. For example, the model training device simulates braking performance data in a second driving environment based on the second braking control data. Furthermore, the model training device adjusts the first parameters of the at least one model to be trained based on the second braking performance data. That is, the model parameters are continuously updated based on the braking performance data, thereby optimizing the braking strategy output by the model and improving the safety and comfort of the braking strategy.

[0154] Of course, the process of inputting training data, obtaining braking performance data, and updating model parameters based on braking performance data can be executed multiple times until the model training is complete.

[0155] Optionally, at least one model can be a single model or multiple models. For example, at least one model may include one or more of an encoder and / or decoder, an inference model, a cue word optimization model, and a cue word optimization model. The encoder provides feature suggestions to the input data, and the extracted features are input into the inference model to obtain the output result. Furthermore, the inference process is guided by cue words to guide the inference model, and the output process can be optimized multiple times. The cue word optimization model is used to continuously optimize the cue words to make the output result of the inference model more accurate.

[0156] In some possible implementations, at least one model includes a machine learning model and / or a neural network model. For example, at least one model includes a first model, which can be a large model, meaning a model with a large number of parameters and a complex structure. More exemplarily, the first model can be a convolutional neural network (CNN) model, a recurrent neural network (RNN) model, or a Transformer-based model, etc. Please refer to Figure 6, which is a schematic diagram of a neural network model provided in an embodiment of this application. The first model is a neural network model including an input layer, hidden layers, and an output layer. Information from multiple feature dimensions in the environmental information can be input into the input layer, and after processing through multiple layers such as the hidden layer or the output layer, the model's output is obtained. It should be understood that the input-output relationship and feature dimensions of the layers shown in Figure 6 are merely examples.

[0157] In some possible implementations, the model training process is supervised by a fault detection module. During model training, if a fault occurs in the braking performance data output by the simulated model, such as a vehicle collision or abnormal vehicle vibration, the model training device can proactively stop the training. Furthermore, the model training device can output prompts indicating abnormal vehicle operation or model training, and can also output the cause of the abnormality to help the user locate the problem.

[0158] As one possible design, the braking control device detects the vehicle's operating state when braking is controlled by second braking control data. When the vehicle's operating state is detected to be normal, the first parameter of at least one model to be trained is adjusted based on the second braking performance data. That is, the process of adjusting model parameters continues even when the vehicle's operating state is normal. This ensures that model training is based on the stable and safe operation of the vehicle, guaranteeing stable model training and improving the model's application safety, while also preventing the model from adjusting in the wrong direction and improving training efficiency.

[0159] In some approaches, during model training and the detection of vehicle operating status, the object being detected may not be an actual vehicle, but rather a simulated vehicle. In this case, the detected data is simulated vehicle operating data.

[0160] In the embodiment shown in Figure 4, the model training device can train at least one model based on the vehicle's driving environment information and its braking performance in these environments. This enables the at least one model to adaptively output optimal braking control data based on the current environment, thereby achieving optimal braking performance. When the model is applied, the user provides a braking trigger instruction (equivalent to turning on a switch), and at least one model can perform braking control based on the current driving scenario information. This model can prevent the driver from applying inappropriate pedal force, which could result in unsafe or inappropriate braking distance and acceleration.

[0161] The braking control method of this application is described below. Please refer to Figure 7, which is a schematic flowchart of a braking control method provided in an embodiment of this application. Optionally, the method is applied to a control device, such as the control device 14 shown in Figure 1. For ease of description, the following description uses a braking control device as the executing entity.

[0162] The braking control method shown in Figure 7 may include one or more steps from S701 to S703. It should be understood that, for ease of description, the steps are described in the order of S701 to S703, but this embodiment does not limit the order of execution, the execution time, or the number of executions of the above steps. S701 to S703 are as follows:

[0163] S701, the braking control device acquires the first driving environment information.

[0164] The first driving environment information describes the environment in which the vehicle is located, and may include information about the vehicle's external environment. Furthermore, the vehicle here may be the vehicle where the braking control device is located or the vehicle specifically serving the vehicle, which differs from the general term "vehicle" used in the aforementioned model training. The vehicle also includes a braking device used to perform braking operations, causing the vehicle to decelerate or cease acceleration.

[0165] The first driving environment information includes information about the vehicle's external environment and / or internal environment. The external environment information includes one or more of the following: weather information, road surface traction indication information, road type, emergency information, or traffic congestion information. The internal environment information includes one or more of the following: driver's seating posture information, driver's seat pressure information, occupant characteristics, cabin detection information, or driver's user characteristics. For a detailed description of the above information, please refer to the preceding explanation of driving environment information; it will not be repeated here.

[0166] In some possible implementations, the first driving environment information includes information indicating at least two feature dimensions, that is, describing the driving environment from at least two dimensions. For example, the first driving environment information includes information on the following six feature dimensions: driver's seating posture information, driver's seat pressure data, weather information, road type, road emergency information, and traffic congestion information.

[0167] Optionally, driving environment information can be obtained through a network and / or by analyzing data collected by the vehicle's perception system.

[0168] Taking weather information as an example, in one example, the vehicle receives weather information through an interface provided by a server. In another example, the perception system includes vision sensors, and the braking control device can analyze the image data collected by the vision sensors to obtain weather information.

[0169] Taking the category of the road where the vehicle is located as an example, in one example, the category of the road where the vehicle is located can be provided by a server or other devices (such as other vehicles or roadside equipment). In another example, the category of the road where the vehicle is located can be obtained through analysis of data collected by the perception system.

[0170] It should be understood that the information mentioned above, which is related to users, is collected or controlled with the user's consent and within the scope of laws and regulations. Furthermore, the aforementioned information is stored, transmitted, and used in accordance with laws and regulations.

[0171] S702, the braking control device obtains the first braking trigger indication input by the first user.

[0172] The first user is a person or program capable of giving a braking trigger instruction.

[0173] In some possible implementations, the first user is the driver of the vehicle. Since the first user can provide brake triggering instructions and control the vehicle's movement and braking, the first user can act as the driver of the vehicle. For example, the first user is a user sitting in the driver's seat of the vehicle.

[0174] In some other possible implementations, the first user logs into the vehicle's infotainment system, meaning the driver's account information in the vehicle is the first user's information. In this case, the brake trigger indication obtained by the brake control is considered to be input by the first user.

[0175] Brake trigger indication includes signals indicating braking, which can be triggered by the brake pedal or other braking trigger methods, such as the parking brake, physical buttons (buttons), user interface, voice triggering, or gesture triggering.

[0176] As a possible example, a vehicle includes a brake pedal and a pedal sensor, which detects brake pedal depressing data. When braking is required, a first user depresses the brake pedal, and the pedal sensor detects this depressing data; therefore, the data collected by the pedal sensor can serve as a brake trigger indication. Accordingly, the brake control unit receives the first brake trigger indication from the pedal sensor. Using brake pedal depressing data as the brake trigger indication aligns with driver braking habits, has low user learning costs, is compatible with most existing vehicle architectures, and offers good applicability.

[0177] As another possible example, the vehicle also includes a brake switch (i.e., a button or key indicating braking), which a first user can press, and the brake switch generates a brake trigger indication in response to the user's press operation. Accordingly, the brake control device receives the first brake trigger indication from the brake switch.

[0178] In some cases, the brake trigger indication may be falsely triggered. For example, referring to Figure 2, if the vehicle 10 includes an accelerator pedal and a brake pedal, and the accelerator pedal and brake pedal are positioned close together, the user may accidentally press the brake pedal when they want to press the accelerator pedal, causing the brake trigger indication to be falsely triggered, resulting in sudden braking of the vehicle and affecting the riding experience.

[0179] In some possible implementations, after receiving the first brake trigger indication, the brake control device can perform false trigger detection on the first brake trigger indication to detect whether the brake trigger indication is falsely triggered, thereby improving braking accuracy. Further, the following step S703 is performed if the first brake trigger indication is not falsely triggered. Four detection methods for determining whether the brake trigger indication is falsely triggered are described below.

[0180] Detection Method 1: The brake control device determines whether the first brake trigger indication is falsely triggered based on brake pedal depress data and the driver's depressing habits. In other words, the brake control device analyzes the brake pedal depress data and the driver's depressing habits to determine if the first brake trigger indication is falsely triggered. For example, if the pressure change curve of the user's current depressor data differs significantly from the pressure curve the user typically uses, then the first brake trigger indication is determined to be falsely triggered.

[0181] Detection Method 2: The braking control device determines whether the first brake trigger indication is falsely triggered based on brake pedal depress data and vehicle driving scenario information. The vehicle driving scenario information includes one or more of the following: obstacle information in the driving scenario, map information or road conditions such as vehicle location (or road), congestion level, and emergency status. Analyzing this vehicle driving scenario information allows for a more accurate determination of whether the brake trigger indication is falsely triggered, thus contributing to improved vehicle driving safety.

[0182] Furthermore, detection method 1 and detection method 2 can be combined. As an example of such a combination, the brake control device determines whether the first brake trigger indication is falsely triggered based on brake pedal depress data, driver's depressing habits data, and vehicle driving scenario information.

[0183] Detection method 3: The braking control device determines whether the first braking trigger indication is falsely triggered based on the number of braking trigger indications received within the second time period. The braking trigger indications received within the second time period include the first braking trigger indication.

[0184] When a user needs to brake, the driver typically triggers the brakes continuously or multiple times. In the above implementation, the number of times the driver triggers the brakes within a continuous time period is monitored to determine whether the driver has accidentally triggered the brakes. For example, if the driver gives two consecutive brake trigger instructions within a limited time (e.g., within 2 seconds, within 1 second, etc.), it is determined that the driver intends to brake, i.e., the brake trigger instructions are not accidental. Optionally, the second time period can be predefined, or the length of the second time period can be related to the vehicle's environment.

[0185] Detection method 4: The brake control device acquires second perception data of the driver of the vehicle after the generation time of the first brake trigger indication, and determines whether the first brake trigger indication is falsely triggered based on the second perception data.

[0186] Understandably, when a driver accidentally triggers the brakes, they may be in a state of tension, panic, excitement, or fear. Combining driver perception information with brake mis-triggering judgment can improve the accuracy of brake mis-triggering detection.

[0187] For example, the second perception data includes one or more of the following: driver's gaze recognition data, driver's fatigue level detection data, driver's attention detection data, driver's facial expression analysis data, driver's image, driver's action analysis, driver's voice data, etc.

[0188] It should be understood that the above-mentioned detection methods can be combined, and the combination will not be explained in detail here.

[0189] S703, the braking control device responds to the first braking trigger indication and obtains first braking control data based on the first driving environment information.

[0190] The first braking control data is used to indicate a braking strategy (braking scheme) and is used to control the braking device. For example, the braking control data includes braking pressure data (i.e., the pressure used for braking), braking acceleration data, etc. The braking control device can use this braking control data to control the braking device and achieve braking of the vehicle. In some possible embodiments, the first braking control data includes pressure data and / or braking acceleration data. The pressure data indicates the brake pedal pressure value within a first time period, where the first time period represents a period of time. In some embodiments, the length of the first time period is not fixed. For example, the pressure data can be a pressure curve, which indicates the braking force at different times within a period of time. The braking acceleration data indicates the braking acceleration within the first time period, where the first time period also indicates a period of time. For example, the braking acceleration data can be a braking acceleration curve, which indicates the braking acceleration at different times within a period of time.

[0191] The braking control system can analyze the vehicle's environment and determine an appropriate braking strategy. When braking, the user only needs to provide a braking trigger instruction, such as lightly pressing the brake pedal, and the system will automatically brake according to the environment. After the user provides the braking trigger instruction (equivalent to turning on a switch), the braking control system can brake based on the current scenario, increasing driving safety and improving the driving experience. For example, if the braking trigger instruction is given by the user pressing the brake pedal, the braking trigger instruction generated by the braking control system is independent of the force with which the brake pedal is pressed. Thus, the user can use their usual force to press the brake pedal, and the system will automatically analyze the driving environment information to obtain appropriate braking control data.

[0192] In some possible implementations, the first braking control data corresponds to a first user. That is, the first braking control data is tailored to the first user. For example, in obtaining the braking control data, information about the first user is utilized, such as the first user's identity or one or more of the first user's user characteristics, including age, gender, clothing, mood, and weight. Different braking control data may be output for other users different from the first user in the same scenario. Furthermore, the parameters of the algorithm or model used to obtain the braking control data are adapted to the first user, or adjusted based on the first user's identity or usage habits. In other words, the algorithm or model used to obtain the braking control data can generate personalized braking control data for the first user based on driving environment information.

[0193] In some possible implementations, the first braking control data is generated by a model, or calculated by an algorithm. Two possible implementations are described below:

[0194] To achieve step 1, please refer to Figure 8. The braking control device inputs first driving environment information into at least one model to obtain first braking control data. Further, at least one model is trained based on the driving environment information and the corresponding braking performance data. The model learns from historical data (i.e., training data), analyzes the correlations and patterns in the historical data, and performs tasks such as prediction or classification. Since driving environment information includes multi-dimensional data and may contain various difficult examples, using the model to infer scenario-adaptive braking control data ensures optimal braking acceleration and distance when dealing with complex driving environments, increasing driving safety and improving the driving experience. Furthermore, the model has strong generalization ability; using the model to obtain braking control data improves the decision-making accuracy of braking control data, thereby enhancing braking safety.

[0195] Optionally, the at least one model can be trained using the model training method shown in Figure 4. Further, the at least one model includes a machine learning model and / or a neural network model. For example, the at least one model includes a first model, which is a large model, meaning a model with a large number of parameters and a complex structure. Exemplarily, the first model can be a CNN model, an RNN model, or a Transformer-based model, etc.

[0196] In some cases, the model can be deployed in separate computing units that are adapted to the model's computational characteristics. In this scenario, obtaining braking control data through the model allows for the rational allocation of computing resources, improving resource utilization.

[0197] Implementation 2: The braking control device obtains first braking control data based on the first driving environment information using a predefined algorithm. For example, using a rule-based algorithm, the first braking control data is output when the first driving environment information meets a first condition. For instance, if there are no obstacles ahead and the weather is not rainy or snowy, the device first decelerates with acceleration A1 for a duration T1, and then decelerates with acceleration A2 for a duration T2, where A1 and A2 are used to exemplarily represent accelerations, and T1 and T2 are used to exemplarily represent durations.

[0198] The process of generating braking control data has been described above as an example. Other possible implementations of the embodiments of this application are described below. It should be understood that the various possible implementations described in this application can be combined.

[0199] In some other possible implementations, the vehicle includes a braking device. As shown in FIG8, after obtaining first braking control data, the braking control device can control the braking device to brake based on the first braking control data. Exemplarily, the braking control device has the capability to control the braking device and can control the braking device to perform braking based on the braking control data. The braking control device outputs a control signal to the braking device, which is generated based on the first braking control data. Again, exemplarily, the control device outputs the first braking control data to the braking system, and the controller within the braking system generates a control signal based on the first braking control data to control the actuators in the braking system to perform braking.

[0200] Alternatively, the braking device may include, for example, a brake motor, a brake pump, or a brake transmission device.

[0201] In some other possible implementations, where the braking control data is obtained using a model, after the braking device is controlled by the first braking control device, the braking control device senses the user's feedback on the braking control data and updates the model's parameters based on the user's feedback; this is adaptive training, or fine-tuning training. Adaptive training allows the model's parameters to be adjusted individually for the user. These adjustments are usually minor, i.e., fine-tuning. Through adaptive training, the model can be continuously improved, ensuring that the model's subsequent outputs conform to the user's usage habits. The improved model incorporates user habits and preferences, is more adapted to the user's braking habits, and enhances the user experience.

[0202] As one possible implementation, the braking control device controls the braking system based on first braking control data and acquires first perception data of the first user of the vehicle during a second time period after braking begins. Based on the first perception data, the braking control device updates first model parameters of at least one model, which are associated with the first user. Optionally, the perception data includes data from one or more dimensions, such as the pressure exerted by the user on the seat, the user's facial expression, the user's voice, and the user's emotions.

[0203] Optionally, when updating the first model parameters of at least one model, if there are multiple models within the at least one model, the braking control device may only update the parameters of some models within the at least one model, while leaving the parameters of other models unchanged. Of course, this application also applies to the case where the model parameters of all models are updated.

[0204] For example, the sensing data includes pressure data collected by at least one pressure sensor located at the driver's seat of the vehicle. When the fluctuation of the pressure data meets a preset fluctuation condition, the braking control device adjusts the first model parameters of at least one model in the convergence direction. And / or, when the fluctuation of the pressure data does not meet a preset fluctuation condition, the first model parameters of at least one model are adjusted in the opposite direction of the convergence direction.

[0205] Taking the pressure information from the seat pressure sensor as a feedback signal during model adjustment as an example, in the adaptive training process, the braking control data output by the model is used to control the braking device. During braking, if the parameters of the seat pressure sensor are stable (i.e., the pressure data meets the fluctuation condition), it indicates that the driver is adapting to the braking strategy output by the current model, and the model parameter adjustment proceeds in the direction of fitting convergence. Conversely, if the seat pressure sensor parameters change significantly (e.g., the parameters of sensors in the leg rest and lumbar support fluctuate greatly), it indicates that the driver is not adapting to the braking strategy output by the current model and is therefore attempting to press the brake pedal to change the braking strategy; in this case, the model parameter adjustment proceeds in the opposite direction of fitting convergence.

[0206] In some cases, the process of outputting braking control data and then adjusting model parameters based on user perception data can be executed multiple times. This allows the model to undergo multiple iterations to obtain an optimal model that incorporates user preferences. Furthermore, after model training is complete, the process of obtaining braking control data based on the driving environment can continue. The relevant steps are described as above, the difference being that it is no longer necessary to update the model parameters based on the driver's feedback on braking operations.

[0207] Furthermore, the model parameters after adaptive training are related to the driver account logged in during adaptive training. The model parameters after adaptive training can be associated with the driver account, for example, stored in the corresponding driver account.

[0208] Furthermore, after adaptive training, at least one model can continue to respond to braking instructions given by the user and obtain corresponding braking control data based on the currently acquired driving environment information.

[0209] In some other possible implementations, during adaptive training, the braking control device can detect the vehicle's operating status. If the vehicle is operating normally, the model parameter adjustment process is executed normally; if the vehicle's operation is abnormal, the model parameter adjustment is stopped. For example, the braking control device detects the vehicle's operating status. If the vehicle is operating normally, the step of "updating the first model parameters of at least one model based on the first perception data" is executed. Alternatively, if the vehicle is operating abnormally, the step of "updating the first model parameters of at least one model based on the first perception data" is not executed. Abnormal operation could be, for example, a vehicle collision or a malfunction in the vehicle's braking system.

[0210] In some other possible implementations, before executing step S703, the first user has logged into the vehicle or the vehicle's infotainment system, and the parameters used by at least one model are those corresponding to the first user. For example, before executing step S703, the braking control device obtains the first user's identification information and determines the first model parameters applied by at least one model based on this information. These first model parameters are associated with the first user. In other words, the braking control device sets the model parameters applied by at least one model to the first model parameters corresponding to the first user based on the first user's identification information. Thus, the braking control data output by at least one model can match the first user's braking preferences, improving the first user's braking experience.

[0211] In some other possible implementations, step S703 is performed when the first function is activated. As one possible design, the braking control device, in response to a first braking trigger indication, obtains first braking control data based on first driving environment information when the first function is activated. It should be understood that the name of this first function can be, for example, adaptive braking function, automatic braking function, scenario braking, etc., and can be flexibly designed.

[0212] As another possible design, before obtaining the first braking control data based on the first driving environment information, the method further includes: receiving a first operation instruction input by a user, the first operation instruction being used to instruct the activation of a first function.

[0213] As another possible design, the method also includes: outputting a first prompt message, which indicates that the first function is active. The first prompt message can be output in the form of interface, voice, light, vibration, etc.

[0214] In some other possible implementations, the braking control device deactivates the first function, i.e., exits the first function, when it detects an abnormal vehicle operating state. Alternatively, when it detects an abnormal vehicle operating state, it stops performing the operation of "obtaining braking control data based on driving environment information". For example, in a scenario where braking control data is obtained using a model, at least one model is stopped from being used when an abnormal vehicle operating state is detected.

[0215] In some solutions, the braking control unit analyzes vehicle driving data to determine the vehicle's operating status. As one possible design, the braking control unit acquires vehicle driving data and, based on this data, deactivates at least one model. That is, the braking control unit can analyze vehicle driving data and manage model usage based on this data. For example, it can promptly disable models when vehicle operation is abnormal, thereby making model application safer and improving vehicle safety and operational stability.

[0216] The above methods and possible implementations can be combined. Furthermore, the above methods and possible implementations may also be combined with existing solutions. For example, in the case of adaptive braking, the stiffness of the brake pedal can be determined by the user or adaptively, so that the user can use a more appropriate force when pressing the brake pedal.

[0217] In the embodiment shown in Figure 7, the braking control device can respond to the braking instruction triggered by the user and determine the braking strategy for the user based on the current driving scenario. The user only needs to give the braking trigger command, and the vehicle can brake smoothly and safely, improving the user's braking experience and enhancing the vehicle's intelligence level. Furthermore, the braking strategy can be personalized and associated with the user who triggers the braking, so the braking strategy can adapt to the user's braking needs, further improving the user experience.

[0218] The foregoing has described the application scenarios and methods provided by the embodiments of this application. The apparatus of the embodiments of this application is provided below. It is understood that the various apparatuses provided in the embodiments of this application, such as interactive devices, computing devices, chips, etc., include hardware structures, software units, or combinations of hardware and software structures to perform the functions described in the above method embodiments. Those skilled in the art should readily recognize that the apparatus and modules within it can be implemented in hardware or a combination of hardware and computer software in conjunction with the various functions described in the embodiments disclosed herein. Whether a function is executed by hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different apparatus implementations in different application scenarios to implement the aforementioned method embodiments, and different implementations of the apparatus should not be considered beyond the scope of the embodiments of this application.

[0219] Several possible devices are listed below.

[0220] Please refer to Figure 9, which is a schematic diagram of a braking control device provided in an embodiment of this application. The braking control device 90 includes a first acquisition unit 901 and a processing unit 902. The braking control device 90 can be an independent device, such as the control device 14 shown in Figure 1. Alternatively, the braking control device 90 can also be a software module and / or hardware module in an independent device, such as a chip in the control device 14.

[0221] The braking control device 90 is used to implement the aforementioned braking control method, such as the method executed by the braking control device in the embodiment shown in FIG7 and its possible implementations. The first acquisition unit 901 is used to acquire data, for example, to perform operations such as acquisition, reception, and collection. The processing unit 902 is used to process data and perform operations, for example, to implement one or more of the aforementioned operations such as data processing, model running, execution, and control. This braking control device is used to implement the method described in the first aspect or any possible implementation of the first aspect.

[0222] In one possible implementation, the first acquisition unit 901 is used to acquire first driving environment information and a first brake trigger indication input by a first user. The processing unit 902 is used to obtain first brake control data based on the first driving environment information in response to the first brake trigger indication. Related descriptions can be found in the embodiment shown in FIG7 above.

[0223] In one possible implementation, the processing unit 902 is used to input first driving environment information into at least one model to obtain first braking control data. Further, the at least one model is trained based on the driving environment information and the corresponding braking performance data.

[0224] In one possible implementation, the processing unit 902 is used to obtain first braking control data based on first driving environment information using a predefined algorithm.

[0225] In one possible implementation, the vehicle further includes a brake pedal and a pedal sensor, the pedal sensor being used to detect brake pedal depressing data. A first acquisition unit 901 is used to receive a first brake trigger indication from the pedal sensor, the first brake control data being independent of the brake pedal depressing force.

[0226] In one possible implementation, the vehicle further includes a brake switch (i.e., a button or key indicating braking), which a first user can press to generate a brake trigger indication in response to the user's press operation. A first acquisition unit 901 is configured to receive the first brake trigger indication from the brake switch.

[0227] In one possible implementation, the processing unit 902 is also used to control the braking device to brake based on the first braking control data.

[0228] In one possible implementation, the processing unit 902 is further configured to control the braking device to brake based on the first braking control data, and the first acquisition unit 901 is further configured to acquire the first perception data of the first user of the vehicle during a second time period after the braking begins. The processing unit 902 is further configured to update the first model parameters of at least one model based on the first perception data, the first model parameters being associated with the first user.

[0229] In one possible implementation, the sensing data includes pressure data collected by at least one pressure sensor located at the driver's seat of the vehicle. The processing unit 902 is further configured to adjust the first model parameters of at least one model in the convergence direction when the fluctuation of the pressure data meets a preset fluctuation condition. And / or, the processing unit 902 is further configured to adjust the first model parameters of at least one model in the opposite direction of the convergence direction when the fluctuation of the pressure data does not meet a preset fluctuation condition.

[0230] In one possible implementation, the processing unit 902 is further configured to detect the vehicle's operating status. If the vehicle is operating normally, the step of "updating the first model parameters of at least one model based on the first perception data" is executed. Alternatively, if the vehicle is operating abnormally, the step of "updating the first model parameters of at least one model based on the first perception data" is not executed.

[0231] In one possible implementation, the first acquisition unit 901 is further configured to acquire the identity information of the first user of the vehicle, and the processing unit 902 is further configured to determine, based on the identity information of the first user of the vehicle, a first model parameter applied by at least one model, wherein the first model parameter is associated with the first user.

[0232] In one possible implementation, the processing unit 902 is also used to detect whether the first brake trigger indication is a false trigger.

[0233] In one possible implementation, the processing unit 902 is used to obtain first braking control data based on first driving environment information when the first braking trigger indication is not a false trigger.

[0234] In one possible implementation, the processing unit 902 is further configured to determine whether the first brake trigger indication is a false trigger based on the brake pedal depress data and the driver's depressing habits data.

[0235] In one possible implementation, the processing unit 902 is further configured to determine whether the first brake trigger indication is a false trigger based on the brake pedal depress data and vehicle driving scenario information.

[0236] In one possible implementation, the processing unit 902 is further configured to determine whether the first brake trigger indication is a false trigger based on brake pedal depress data, driver's depressing habits data, and vehicle driving scenario information.

[0237] In one possible implementation, the processing unit 902 is further configured to determine whether the first brake trigger indication is a false trigger based on the number of brake trigger indications received within the second time period, wherein the brake trigger indications received within the second time period include the first brake trigger indication.

[0238] In one possible implementation, the vehicle further includes a driver monitoring system. The first acquisition unit 901 is further configured to acquire second perception data of the driver of the vehicle after the generation time of the first brake trigger indication, and the processing unit 902 is further configured to determine, based on the second perception data, whether the first brake trigger indication was falsely triggered.

[0239] As another possible design, the first acquisition unit 901 is also used to receive a first operation instruction input by the user, the first operation instruction being used to indicate the activation of a first function.

[0240] As another possible design, the braking control device also includes an output unit (not shown in the figure), which outputs a first prompt message to indicate that the first function is active. The first prompt message can be output in the form of an interface, voice, light, vibration, etc.

[0241] In one possible implementation, the processing unit 902 is further configured to deactivate the first function, i.e., exit the first function, when an abnormal operating state of the vehicle is detected. Alternatively, when an abnormal operating state of the vehicle is detected, the operation of "obtaining braking control data based on driving environment information" is no longer performed.

[0242] Please refer to Figure 10, which is a schematic diagram of a model training device provided in an embodiment of this application. The model training device 100 includes a second acquisition unit 1001 and a training unit 1002. The model training device 100 can be a standalone device, such as a computing device (e.g., a host or server). Alternatively, the model training device 100 can also be a software module and / or hardware module in a standalone device, such as a chip in a computing device.

[0243] The model training device 100 is used to implement the aforementioned model training method, such as the method and its possible implementations described in the embodiment shown in FIG4. The second acquisition unit 1001 is used to acquire data, for example, to perform operations such as acquisition, reception, and collection. The training unit 1002 is used to process data and perform operations, for example, to implement one or more of the aforementioned operations such as data processing, model running, training, execution, and control.

[0244] In one possible implementation, the second acquisition unit 1001 is used to acquire multiple training data, and the training unit 1002 is used to train at least one model based on the multiple training data. The at least one model is used to obtain target braking control data based on the current driving environment information in response to a braking trigger indication. The target braking control data is used to control the braking device of the vehicle.

[0245] In another possible implementation, the training unit 1002 is used to input the second driving environment information into at least one model to be trained to obtain the second braking control data, acquire the second braking performance data when the vehicle is controlled to brake with the second braking control data under the second driving environment information, and adjust the first parameter of the at least one model to be trained based on the second braking performance data.

[0246] In another possible implementation, the training unit 1002 is further configured to detect the vehicle's operating state when the vehicle is controlled to brake with the second braking control data, and when the vehicle's operating state is detected to be normal, adjust the first parameter of at least one model to be trained based on the second braking performance data.

[0247] Figure 11 shows a schematic diagram of a control device provided in an embodiment of this application. The control device 14 is a device with computing capabilities; this device can be a physical device, such as an embedded device. Optionally, the control device 14 can be included in a vehicle, for example, as an on-board component in the vehicle 10 shown in Figure 2.

[0248] As shown in Figure 11, the control device 14 includes a processor 141 and a memory 142. Optionally, the control device 14 may also include one or more of a connection line 144, a communication interface 143, etc. For example, the processor 141 and the memory 142 communicate with each other via the connection line 144. It should be understood that this application does not limit the number of processors and memories in the control device 14.

[0249] Memory 142 provides storage space for computer programs or data. Memory 142 may include volatile memory, such as random access memory (RAM). Memory 142 may also include non-volatile memory, such as read-only memory (ROM), flash memory, hard disk drive (HDD), or solid state drive (SSD).

[0250] Processor 141 is a module for performing calculations and may include any one or more of the following: controller, central processing unit (CPU), graphics processing unit (GPU), microprocessor (MP), digital signal processor (DSP), coprocessor (to assist the central processing unit in completing corresponding processing and applications), application-specific integrated circuit (ASIC), microcontroller unit (MCU), virtual machine, container, etc.

[0251] The communication interface 143 is used to provide information input or output to at least one processor, such as an in-line interface, an out-of-line interface, etc. And / or, the communication interface 143 can be used to receive externally transmitted data and / or transmit data externally. The communication interface 143 can be a wired link interface, including an Ethernet cable, or a wireless link interface (Wi-Fi, Bluetooth, general wireless transmission, and other wireless communication technologies, etc.). Optionally, the communication interface 143 may also include a transmitter (such as a radio frequency transmitter, antenna, etc.) or a receiver coupled to the interface.

[0252] The connection line 144 can be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, only one line is used in Figure 11, but this does not imply that there is only one bus or one type of bus. The connection line 144 can include pathways for transmitting information between various components of the control device 14 (e.g., memory 142, processor 141, communication interface 143).

[0253] In one possible implementation, memory 142 stores executable instructions, and processor 141 executes the executable instructions to implement the aforementioned braking control method, such as the method executed by the braking control device in the embodiment shown in FIG7 and its possible implementation methods.

[0254] Figure 12 shows a schematic diagram of a computing device provided in an embodiment of this application. The computing device 120 is a device with computing capabilities, and the device can be a physical device, such as an embedded device. Optionally, the computing device 120 can be included in a vehicle, such as an on-board component in the vehicle 10 shown in Figure 2.

[0255] As shown in Figure 12, the computing device 120 includes a processor 1201 and a memory 1202. Optionally, the computing device 120 may also include one or more of a connection line 1204, a communication interface 1203, etc. For example, the processor 1201 and the memory 1202 communicate with each other via the connection line 1204. It should be understood that this application does not limit the number of processors and memories in the computing device 120.

[0256] Memory 1202 provides storage space for computer programs or data. Memory 1202 may include volatile memory, such as RAM. Memory 1202 may also include non-volatile memory, such as ROM, flash memory, HDD, or SSD.

[0257] Processor 1201 is a module for performing calculations and may include any one or more of the following: controller, CPU, GPU, MP, DSP, coprocessor, ASIC, MCU, virtual machine, container, etc.

[0258] The communication interface 1203 is used to provide information input or output to at least one processor, such as an in-line interface, an out-of-line interface, etc. And / or, the communication interface 1203 can be used to receive externally transmitted data and / or transmit data externally. The communication interface 1203 can be a wired link interface including an Ethernet cable, or a wireless link interface. Optionally, the communication interface 1203 may also include a transmitter or receiver coupled to the interface.

[0259] The connection line 1204 can be a PCI bus or an EISA bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, only one line is used in Figure 12, but this does not imply that there is only one bus or one type of bus. The connection line 1204 can include a path for transmitting information between various components of the computing device 120 (e.g., memory 1202, processor 1201, communication interface 1203).

[0260] In one possible implementation, the memory 1202 stores executable instructions, and the processor 1201 executes the executable instructions to implement the aforementioned model training method, such as the method executed by the model training device in the embodiment shown in FIG4 and its possible implementation methods.

[0261] This application also provides a chip, including a processor and a communication interface. The communication interface is used for outputting and / or outputting data (including instructions), and / or for receiving and / or sending data. When the processor executes program instructions in memory, it performs the aforementioned method for updating and upgrading software, for example, implementing the method shown in the embodiments of FIG4 or FIG7.

[0262] This application also provides a braking system, including a braking device 12, and a control device 14 or a braking control device 90.

[0263] This application provides a computer-readable storage medium storing instructions that, when executed by at least one processor, implement the aforementioned method for updating and upgrading software, such as the method shown in the embodiments of FIG4 or FIG7. The computer-readable storage medium can be any available medium capable of being stored by a computing device, or a data storage device such as a data center containing one or more available media. The computer-readable storage medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state drive).

[0264] This application provides a computer program product including computer instructions that, when executed on at least one processor, implement the aforementioned method for updating and upgrading software, such as the method in the embodiments shown in FIG4 or FIG7. Optionally, the computer program product may be a software installation package or image package. When the aforementioned method is required, the computer program product can be downloaded and executed on a computing device.

[0265] This application provides a vehicle that includes the aforementioned brake control device 90, or the aforementioned control device 14, or the aforementioned braking system, or the aforementioned chip, or the aforementioned computer storage medium, or the aforementioned computer program product.

[0266] In addition, a few additional points need to be made regarding this application:

[0267] 1. Unless otherwise stated, “multiple” means two or more.

[0268] 2. Unless otherwise specified or in case of logical conflict, the terms and / or descriptions in different embodiments of this application are consistent and can be referenced in each other. The technical features in different embodiments can be combined to form new embodiments according to their inherent logical relationships.

[0269] III. The various numerical designations used in this application are merely for descriptive convenience and are not intended to limit the scope of protection of this application. The magnitude of the serial numbers used in this application does not imply a sequential order of execution; the execution order of each process should be determined by its function and internal logic. For example, the terms "first," "second," "third," "fourth," and other various terminology (if present) in the specification, claims, and drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. Such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein.

[0270] Furthermore, any embodiment or design described in this application as "exemplary" or "for example" should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of terms such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner for ease of understanding.

[0271] IV. The terms “comprising” and “having” and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are expressly listed, but may include other steps or units that are not expressly listed or that are inherent to such process, method, product or device.

[0272] V. In this application, "for indicating" can be understood as "enabling". "Enabling" can include direct enabling and indirect enabling. When describing information for enabling A, it can include whether the information directly enables A or indirectly enables A, but it does not mean that the information necessarily carries A.

[0273] The information that enables the information is called the information to be enabled. In the specific implementation process, there are many ways to enable the information to be enabled, such as, but not limited to, directly enabling the information to be enabled, such as the information to be enabled itself or its index. It can also be indirectly enabled by enabling other information, where there is a relationship between the other information and the information to be enabled. It can also enable only a part of the information to be enabled, while the other parts are known or pre-agreed upon. For example, enabling specific information can be achieved by using a pre-agreed (e.g., protocol-defined) arrangement of various pieces of information, thereby reducing enabling overhead to some extent. Simultaneously, common parts of various pieces of information can be identified and enabled uniformly to reduce the enabling overhead caused by individually enabling the same information.

[0274] VI. In this application, "predefined" may include preconfiguration. For example, predefining certain information means that the information is calculated or received in advance before performing an action that uses the information. The "predefined" can be implemented by pre-storing corresponding codes, tables, or other means that can be used to indicate relevant information in the device (e.g., in a controller or vehicle). This application does not limit the specific implementation method.

[0275] VII. The term "storage" or "preservation" in this application can refer to storage in one or more memory devices. These memory devices can be separately configured or integrated into an encoder, decoder, processor, or communication device. Alternatively, some memory devices can be separately configured, while others can be integrated into a decoder, processor, or communication device. The type of memory can be any form of storage medium, and this is not limited.

[0276] 8. In the schematic diagrams in the accompanying drawings of this application, the dashed arrows or boxes indicate optional steps or optional modules.

[0277] 9. Unless otherwise stated, " / " indicates that the objects before and after are in an "or" relationship. For example, A / B can mean A or B. In this application, "and / or" is merely a description of the relationship between the related objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist simultaneously, and B exists alone. A and B can be singular or plural.

Claims

1. A brake control method characterized by, The method is applied to a vehicle comprising a braking device, and the method further comprises: obtaining first driving environment information, the first driving environment information comprising information of an external environment of the vehicle; obtaining a first brake trigger indication input by a first user, the first brake trigger indication being used to indicate a brake control; obtaining first brake control data based on the first driving environment information in response to the first brake trigger indication, the first brake control data being used to control the braking device to brake, the first brake data corresponding to the first user.

2. The method of claim 1, wherein, The obtaining of the brake control data based on the first driving environment information comprises: inputting the first driving environment information into at least one model to obtain the first brake control data, the at least one model being trained based on driving environment information and brake performance data corresponding to the driving environment information.

3. The method according to claim 1 or 2, characterized in that, The information of the external environment of the vehicle comprises one or more of the following: weather information, a category of a road where the vehicle is located, emergency information of the road where the vehicle is located, or congestion information of the road where the vehicle is located.

4. The method according to any one of claims 1 to 3, characterized in that, The first driving environment information further comprises information of an internal environment of the vehicle, The information of the internal environment of the vehicle comprises one or more of the following:

5. The method according to any one of claims 1 to 4, characterized in that, sitting posture information of a driver of the vehicle or seat pressure information of a driver seat of the vehicle. The obtaining of the first brake trigger indication input by the user comprises:

6. The method according to any one of claims 1 to 5, characterized in that, receiving the first brake trigger indication from a pedal sensor, the first brake control data being irrelevant to a degree of depression of a brake pedal, and the pedal sensor being used to detect depression data of the brake pedal. The first brake control data comprises pressure data and / or brake acceleration data, The pressure data is used to indicate a brake pedal pressure value in a first time period, 7. The method according to any one of claims 1 to 6, characterized in that, The brake acceleration data is used to indicate a brake acceleration in the first time period. The method further comprises:

8. The method of claim 2, wherein, controlling the braking device to brake based on the first brake control data. The method further comprises: controlling the braking device to brake based on the first brake control data; obtaining first perception data of the first user of the vehicle in a second time period after the brake starts; 9. The method of claim 8, wherein, updating first model parameters of the at least one model based on the first perception data, the first model parameters being associated with the first user. The perception data comprises pressure data collected by at least one pressure sensor arranged at a driver seat of the vehicle, and the updating of the first model parameters of the at least one model based on the first perception data comprises: adjusting the first model parameters of the at least one model in a converging direction when fluctuations of the pressure data satisfy a preset fluctuation condition; 10. The method according to any one of claims 2, 8, 9, characterized in that, or adjusting the first model parameters of the at least one model in a direction opposite to the converging direction when fluctuations of the pressure data do not satisfy the preset fluctuation condition. Before the inputting of the first driving environment information into the at least one model to obtain the first brake control data, the method further comprises: obtaining identity information of the first user of the vehicle. determine, based on the identity information of the first user of the vehicle, a first model parameter applied by the at least one model, the first model parameter being associated with the first user.

11. The method according to any one of claims 1 to 10, characterized in that, After obtaining the first brake trigger indication input by the user, the method further comprises: detecting whether the first brake trigger indication is a false trigger; the first brake control data based on the first driving environment information comprises: in a case where the first brake trigger indication is not a false trigger, obtaining first brake control data based on the first driving environment information.

12. The method of claim 11, wherein the detection of whether the first brake trigger indication is a false trigger comprises: determining, based on the pedal data of the brake pedal and the pedal habit data of the driver of the vehicle, whether the first brake trigger indication is a false trigger.

13. The method of claim 11, wherein, the detection of whether the first brake trigger indication is a false trigger comprises: determining, based on the pedal data of the brake pedal, the pedal habit data of the driver of the vehicle, and the vehicle driving scene information, whether the first brake trigger indication is a false trigger.

14. The method according to any one of claims 11-13, characterized in that, the detection of whether the first brake trigger indication is a false trigger comprises: determining, based on the number of brake trigger indications received within a second time period, whether the first brake trigger indication is a false trigger, the brake trigger indications received within the second time period including the first brake trigger indication.

15. The method according to any one of claims 11-14, characterized in that, the vehicle further comprises a driver monitoring system, and the detection of whether the first brake trigger indication is a false trigger comprises: obtaining second perception data of the driver of the vehicle after the generation time of the first brake trigger indication; determining, based on the second perception data, whether the first brake trigger indication is a false trigger.

16. The method of claim 2, wherein, The method further comprises: obtaining driving data of the vehicle; stopping enabling the at least one model based on the driving data of the vehicle.

17. A model training method, comprising: The method comprises: obtaining a plurality of training data, each of the plurality of training data comprising driving environment information and vehicle brake control data corresponding to the driving environment information, the driving environment information comprising information of an external environment of the vehicle; training at least one model based on the plurality of training data, the at least one model being configured to obtain target brake control data based on current driving environment information in response to a brake trigger indication, the target brake control data being used to control a brake device of the vehicle.

18. The method of claim 17, wherein, The training of the at least one first model based on the plurality of training data comprises: inputting second driving environment information into the at least one model to be trained to obtain second brake control data, the second driving environment information belonging to one of the plurality of training data; obtaining second brake performance data when the vehicle is braked based on the second brake control data under the second driving environment information; adjusting parameters of the at least one model to be trained based on the second brake performance data.

19. The method of claim 18, wherein, The adjustment of the parameters of the at least one model to be trained based on the second brake performance data comprises: detecting an operating state of the vehicle when the vehicle is braked based on the second brake control data; When detecting that the vehicle operating state is normal, adjusting parameters of the at least one model to be trained based on the second braking performance data.

20. The method according to any one of claims 17-19, characterized by, The braking performance data comprises braking acceleration and / or braking distance.

21. The method according to any one of claims 17-20, characterized by, The information of the external environment of the vehicle comprises one or more of the following: weather information, a category of a road where the vehicle is located, emergency information of the road where the vehicle is located, or congestion information of the road where the vehicle is located.

22. The method according to any one of claims 17-21, characterized by, The first driving environment information further comprises information of an internal environment of the vehicle, The information of the internal environment of the vehicle comprises one or more of the following:

23. The method according to any one of claims 17-22, characterized by, sitting posture information of a driver of the vehicle, or seat pressure information of a driver seat of the vehicle. The first braking control data comprises pressure data and / or braking acceleration data, The pressure data is used to indicate a brake pedal pressure value in a first time period, 24. A brake control device characterized by comprising: The braking acceleration data is used to indicate a braking acceleration in the first time period.

25. A model training apparatus, comprising: The braking control device comprises an acquisition unit and a processing unit, and is configured to implement the method of any one of claims 1-16.

26. A control device characterized by comprising: The model training device comprises an acquisition unit and a processing unit, and is configured to implement the method of any one of claims 17-23.

27. A computing device, comprising: The device comprises a processor and a memory, the memory is configured to store computer instructions, and the processor is configured to invoke the computer instructions to implement the method of any one of claims 1-16.

28. A brake system characterized by, The device comprises a processor and a memory, the memory is configured to store computer instructions, and the processor is configured to invoke the computer instructions to implement the method of any one of claims 17-23.

29. A vehicle characterized by The braking system comprises the control device of claim 26 and a braking device, and the braking device is configured to brake based on the braking control data provided by the control device.

30. A computer-readable storage medium, characterized in that, The vehicle comprises the control device of claim 26, or the braking system of claim 28.

31. A computer program product, characterised in that, The computer readable storage medium is configured to store computer program instructions, and when the computer program instructions are executed by a processor, an apparatus comprising the processor implements the method of any one of claims 1-23. The computer program product comprises computer program instructions, and when the computer program instructions are executed by a processor, an apparatus comprising the processor implements the method of any one of claims 1-23.

Images