Method and device for controlling vehicle, vehicle and storage medium

Abstract

The embodiment of the invention provides a method and device for controlling a vehicle, the vehicle and a storage medium. The method comprises the steps of determining a target driving style of a driver of a vehicle in the driving process of the vehicle; processing the target driving style, the environmental data of the vehicle and the state information of the driver through a pre-trained auxiliary driving model to obtain a target control signal; acquiring an actual control signal output by an intelligent driving domain controller of the vehicle based on the target intelligent driving algorithm; adjusting at least one parameter of the target intelligent driving algorithm based on the actual control signal and the target control signal; and controlling the vehicle to run according to the adjusted parameters. According to the technical scheme provided by the embodiment of the invention, the control of the intelligent driving domain controller on the vehicle can be adapted to the driving styles of the driver in different scenes and different states, and the personalized auxiliary driving experience of the driver is improved.

Description

Technical Field

[0001] This application relates to the field of assisted driving technology, and more specifically, to a method, device, vehicle, and storage medium for controlling a vehicle. Background Art

[0002] Personalized intelligent driving experience is one of the important research directions in the field of intelligent driving, and its core lies in the precise matching of the driver's driving style and vehicle driving control.

[0003] In related technologies, the vehicle collects the driver's behavior data during manual driving, such as the operation signals of the brake pedal, the drive pedal, the steering wheel, the gear shift lever, etc. Based on the above behavior data, the driver's driving style is determined. During intelligent assisted driving, the parameters in the intelligent driving algorithm are set to match the driver's driving style, and then the vehicle is controlled to drive according to the parameters matching the above driving style.

[0004] In the above personalized intelligent driving process, the parameters of the intelligent driving algorithm are fixed values matching the driver's driving style, which is not flexible enough. Summary of the Invention

[0005] Embodiments of this application propose a method, device, vehicle, and storage medium for controlling a vehicle.

[0006] In a first aspect, embodiments of this application provide a method for controlling a vehicle, including: during the driving process of the vehicle, determining the target driving style of the driver of the vehicle; processing the target driving style, the real-time collected environmental data of the vehicle, and the real-time collected state information of the driver through a pre-trained assisted driving model to obtain a target control signal; obtaining the actual control signal output by the intelligent driving domain controller of the vehicle based on the target intelligent driving algorithm; adjusting at least one parameter of the target intelligent driving algorithm based on the actual control signal and the target control signal; and controlling the vehicle to drive according to the adjusted parameters.

[0007] In a second aspect, embodiments of this application provide a device for controlling a vehicle, including: a style determination module for determining the target driving style of the driver of the vehicle during the driving process of the vehicle; a first acquisition module for processing the target driving style, the real-time collected environmental data of the vehicle, and the real-time collected state information of the driver through a pre-trained assisted driving model to obtain a target control signal; a second acquisition module for obtaining the actual control signal output by the intelligent driving domain controller of the vehicle based on the target intelligent driving algorithm; a parameter adjustment module for adjusting at least one parameter of the target intelligent driving algorithm based on the actual control signal and the target control signal; and a vehicle control module for controlling the vehicle to drive according to the adjusted parameters.

[0008] In a third aspect, an embodiment of the present application provides a vehicle, including: a memory; one or more processors coupled to the memory; and one or more applications, wherein the one or more applications are stored in the memory and configured to be executed by the one or more processors, and the one or more applications are configured to execute the method described in the first aspect.

[0009] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, in which computer program instructions are stored, and the computer program instructions can be called by a processor to execute the method described in the first aspect.

[0010] In a fifth aspect, an embodiment of the present application provides a computer program product, which is used to implement the method described in the first aspect when the computer program product is executed.

[0011] Compared with the technical solutions provided by the related art, the technical solution provided by the embodiment of the present application, during the driving process of the vehicle, first determines the driving style of the driver, and then determines the target control signal based on the real-time collected environmental data of the vehicle, the state information of the driver, and the determined driving style, and adjusts at least one parameter of the intelligent driving algorithm according to the target control signal and the actual control signal output by the intelligent driving algorithm, and then controls the vehicle to drive according to the adjusted parameters. In this way, the intelligent driving domain controller can adapt the control of the vehicle to the driving styles of the driver in different scenarios and different states, improving the driver's personalized assisted driving experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] To more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without creative efforts.

[0013] Figure 1 is a schematic diagram of an implementation environment provided by an embodiment of the present application.

[0014] Figure 2 is a flowchart of a method for controlling a vehicle provided by an embodiment of the present application.

[0015] Figure 3 is a flowchart of a method for controlling a vehicle provided by another embodiment of the present application.

[0016] Figure 4 is a flowchart of a training process of a driving style recognition model provided by an embodiment of the present application.

[0017] Figure 5It is a flowchart of the training process of the assisted driving model provided by an embodiment of the present application.

[0018] Figure 6 It is a flowchart of the device for controlling a vehicle provided by an embodiment of the present application.

[0019] Figure 7 It is a structural block diagram of a vehicle provided by an embodiment of the present application. Detailed implementation manners

[0020] The following details the implementation manners of the present application. Examples of the implementation manners are shown in the accompanying drawings, where the same or similar reference numerals represent the same or similar elements or elements with the same or similar functions throughout. The implementation manners described below by referring to the accompanying drawings are exemplary only for explaining the present application and should not be construed as limiting the present application.

[0021] To enable those skilled in the art of the present technology to better understand the solutions of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative efforts fall within the scope of protection of the present application.

[0022] Please refer to Figure 1 , which shows a schematic diagram of the implementation environment shown in an embodiment of the present application. The implementation environment includes a vehicle 100, and the vehicle 100 includes an intelligent driving domain controller and an adaptive controller.

[0023] In the embodiments of the present application, the vehicle 100 includes a pre-trained driving style recognition model and an assisted driving model. The driving style recognition model is used to recognize the driving style of the driver of the vehicle 100. The assisted driving model is used to determine a target control signal based on the above driving style, the environmental data of the vehicle 100, and the status information of the driver. The above driving style recognition model can be trained locally on the vehicle 100 or deployed to the vehicle 100 after being trained in the cloud. The above assisted driving model can be trained locally on the vehicle 100 or deployed to the vehicle 100 after being trained in the cloud. In the embodiments of the present application, only the example where both the driving style recognition model and the assisted driving model are deployed to the vehicle 100 after being trained in the cloud is used for illustration.

[0024] In an embodiment of the present application, the vehicle 100 adjusts at least one parameter of the intelligent driving algorithm based on the target control signal and the control signal output by the intelligent driving domain controller. After that, the intelligent driving domain controller controls the vehicle to travel according to the adjusted parameters. In this way, the control of the intelligent driving domain controller over the vehicle can adapt to the driving styles of the driver in different scenarios and different states, improving the driver's personalized assisted driving experience.

[0025] At least one parameter of the intelligent driving algorithm includes but is not limited to: lateral and longitudinal chassis control calibration parameters, active lane change trigger conditions, maximum execution time for a single lane change, minimum safe collision threshold, maximum driving speed, and so on.

[0026] In some embodiments, the vehicle 100 includes an environmental perception system for collecting environmental data of the vehicle 100. The above environmental data includes but is not limited to: road information of the current driving road of the vehicle 100 (such as road boundaries, road widths, number of lanes), obstacle information around the vehicle 100 (such as the shape, speed, relative position relationship with the vehicle 100 of the obstacle, etc.), traffic information around the vehicle 100 (such as the lighting state of traffic lights, road speed limit information), traffic flow of the current driving road of the vehicle 100, and so on. The environmental perception system includes an image acquisition device and a radar. Among them, the image acquisition device can be a surround view camera installed on the roof of the vehicle 100. The radar can include lidar, millimeter wave radar, ultrasonic radar, and so on.

[0027] In some embodiments, the vehicle 100 includes a state acquisition device for determining the state information of the driver. The state information of the driver can include normal state, fatigue state, nervous state, distracted state, and so on. Optionally, the state acquisition device includes a camera installed inside the vehicle 100. The camera can collect the face image of the driver, and by analyzing the eye region in the above face image, the state information of the driver can be determined. For example, when it is recognized that the driver blinks frequently, it can be determined that the driver is in a fatigue state; for another example, when it is recognized that the driver's pupils dilate, it can be determined that the driver is in a nervous state.

[0028] Please refer to Figure 2 , which shows a flowchart of a method for controlling a vehicle provided by an embodiment of the present application. The method includes the following processes.

[0029] S201, during the driving process of the vehicle, determine the target driving style of the driver.

[0030] The driving style of a driver is obtained by clustering the driver's behavioral data of the vehicle. For example, the driving styles of a driver can include normal, cautious, aggressive, agile, and comfortable. The target driving style is any one of the above. When the driver is in different driving styles, the control strategies for the vehicle usually have obvious differences. For example, in congested traffic conditions, if the driver's driving style is aggressive, the driver is more likely to frequently change lanes and drive among other vehicles. If the driver's driving style is cautious, the driver is more likely to follow the vehicle in front. Therefore, in the process of controlling the vehicle through the intelligent driving algorithm, it is necessary to first determine the driver's driving style and then adjust at least one parameter in the intelligent driving algorithm to match the above driving style.

[0031] The vehicle can determine the target driving style based on a pre-trained driving style recognition model or based on the current road conditions. The above two methods will be described in the following embodiments.

[0032] In some embodiments, when the intelligent driving adaptive adjustment function of the vehicle is in the on state, S201 is executed. The intelligent driving adaptive adjustment function refers to the function of the vehicle to adaptively adjust at least one parameter in the intelligent driving algorithm according to the driver's state, the environmental data of the vehicle, and the driver's driving style during the driving process. Whether the intelligent driving adaptive adjustment function is in the on state or the off state can be set by default by the vehicle or customized by the driver. Further, when the intelligent driving adaptive adjustment function of the vehicle is in the on state, the driving style of the driver is determined every preset time period. The preset time period is set according to experiments or experience. For example, it is 1 minute.

[0033] S202, process the environmental data of the vehicle, the state information of the driver of the vehicle, and the target driving style collected in real time through a pre-trained assisted driving model to obtain a target control signal.

[0034] The assisted driving model is obtained by training a second neural network model based on a second training data set. The second training data set includes multiple groups of second training data, and each group of second training data includes: the environmental data of the vehicle, the state information of the driver, the driving style of the driver, and the behavioral data of the driver. The training process of the assisted driving model will be described in the following embodiments.

[0035] The environmental data of the vehicle includes the road information, obstacle information, and traffic information of the current driving road of the vehicle. The road information includes the road width, road boundary, number of lanes, lane line type, etc. of the current driving road. The obstacle information includes the shape, speed, relative position relationship with the vehicle, etc. of the obstacle. The traffic information includes traffic flow, the lighting state of traffic lights, the indication information of traffic signs, etc.

[0036] The status information of the driver can be any one of the normal state, the tense state, and the fatigued state.

[0037] The target control signal may include: the steering direction and angle of the steering wheel, the speed of the vehicle, the acceleration, the gear control signal, the brake control signal, and so on. The signal type of the target control signal can be determined according to the actual assisted driving function, and the embodiments of the present application do not limit this.

[0038] The vehicle inputs the environment data of the vehicle, the status information of the driver of the vehicle, and the target driving style collected in real time into the assisted driving model, and the assisted driving model outputs the target control signal.

[0039] In some embodiments, when the duration ratio of the vehicle in the same driving style within the current time window is greater than the preset ratio, and this driving style is different from the driving style adopted in the previous parameter adjustment process, S202 is executed. The preset ratio is set according to experiments or experience. For example, 60%. Exemplarily, the time window is 10 minutes, the driving style used by the vehicle in the previous parameter adjustment process is the cautious style, however, the time ratio of the vehicle in the aggressive style within the current time window is greater than 60%, then S202 is executed.

[0040] In other embodiments, when the consecutive number of times the vehicle detects that the driver is in the same driving style is greater than the preset number of times, and this driving style is different from the driving style adopted in the previous parameter adjustment process, S202 is executed. The preset number of times is set according to experiments or experience. For example, 3 times.

[0041] In other embodiments, when the vehicle obtains the intelligent driving parameter adjustment instruction, S202 is executed. If the driver determines that the driving process of the vehicle does not conform to his own driving style, the intelligent driving parameter adjustment instruction can be triggered to trigger the adjustment process of at least one parameter of the intelligent driving algorithm. The intelligent driving parameter adjustment instruction can be a voice instruction, a touch instruction on the intelligent driving parameter adjustment control on the touch panel of the vehicle, and so on.

[0042] In other embodiments, when the vehicle monitors that the road condition of the current driving road has changed significantly, S201 is executed. The significant change in the road condition of the vehicle's current driving road may mean that the type of the current driving road has changed, for example, switched from a highway to an urban road. The significant change in the road condition of the vehicle's current driving road may also be that the traffic flow of the current driving road has changed, for example, switched from severe congestion to smooth traffic.

[0043] In addition, after the vehicle is powered on and started, S202 is directly executed when the driving style of the driver is obtained for the first time.

[0044] S202, Obtain the actual control signal output by the intelligent driving domain controller based on the target intelligent driving algorithm.

[0045] The target intelligent driving algorithm is the intelligent driving algorithm currently used by the intelligent driving domain controller. The above actual control signals may include the vehicle's speed, acceleration, the rotation direction and angle of the steering wheel, braking control signals, driving control signals, and so on. The vehicle can directly read the above actual control signals from the local cache.

[0046] S204, Based on the actual control signal and the target control signal, adjust at least one parameter of the intelligent driving algorithm in the vehicle.

[0047] At least one parameter of the intelligent driving algorithm includes but is not limited to: longitudinal and lateral chassis control calibration parameters, active lane change trigger conditions, maximum execution time for a single lane change, minimum safe collision threshold, maximum driving speed, and so on. The adjustment process will be elaborated in the following embodiments.

[0048] S205, Control the vehicle to travel according to the adjusted parameters.

[0049] The intelligent driving domain controller generates a control signal for the vehicle based on the adjusted parameters, and the vehicle travels according to the above control signal.

[0050] In summary, the technical solution provided by the embodiments of the present application, during the driving process of the vehicle, first determines the driving style of the driver, then determines the target control signal based on the real-time collected environmental data of the vehicle, the state information of the driver, and the determined driving style, and adjusts at least one parameter of the intelligent driving algorithm according to the target control signal and the actual control signal output by the intelligent driving algorithm, and then controls the vehicle to travel according to the adjusted parameters. In this way, the control of the intelligent driving domain controller on the vehicle can adapt to the driving styles of the driver in different scenarios and different states, improving the personalized assisted driving experience of the driver.

[0051] Please refer to Figure 3 , which shows the flowchart of the method for controlling a vehicle shown in an embodiment of the present application. In the optional embodiment provided based on Figure 2 the embodiment, S201 is replaced by S301 - S303, S204 is replaced by S306 - S307, and S205 is replaced by S308. The method includes the following process.

[0052] S301, During the driving process of the vehicle, collect environmental data.

[0053] The environmental data of the vehicle includes road information, obstacle information, and traffic information of the current driving road of the vehicle.

[0054] The road information includes the road width, road boundaries, number of lanes, lane line types, etc. of the currently traveled road. Optionally, an image acquisition device provided on the vehicle acquires an environmental image, and the environmental image is subjected to image recognition to obtain the above-mentioned road information. Alternatively, the vehicle obtains its own position information and queries the road information corresponding to the above-mentioned position information according to a pre-stored high-precision map.

[0055] The obstacle information includes the shape, speed, relative position relationship with the vehicle, etc. of the obstacle. Optionally, the above-mentioned obstacle information is detected by a radar on the vehicle.

[0056] The traffic information includes traffic flow, the lighting state of traffic lights, the indication information of traffic signs, etc. Among them, the vehicle can send a traffic flow acquisition request to the server and receive the traffic flow returned by the server according to the traffic flow acquisition request. In addition, the vehicle can also perform recognition on the above-mentioned environmental image to obtain the lighting state of traffic lights and the indication information of traffic signs.

[0057] S302, collect the driver's status information.

[0058] Optionally, the vehicle includes a status acquisition device. The status acquisition device acquires a video containing the driver's face image, and then performs recognition on each frame image in the above-mentioned video. When it is recognized that the driver's blink frequency is greater than a preset frequency, it is determined that the driver is in a fatigued state. When it is recognized that the pupil area of the driver becomes significantly larger, it is determined that the driver is in a tense state. Otherwise, the driver is in a normal state.

[0059] S303, process the environmental data and status information through a pre-trained driving style recognition model to obtain a target driving style.

[0060] The vehicle inputs the environmental data and status information into the driving style recognition model, and the driving style recognition model outputs the target driving style.

[0061] In other possible implementation manners, the vehicle collects the environmental data of the vehicle; determines the driving road condition of the vehicle based on the environmental data of the vehicle; obtains the driving style corresponding to the driving road condition of the vehicle as the target driving style.

[0062] Optionally, the driving conditions of the vehicle are evaluated from two dimensions: road type and traffic flow. For example, the road type includes four states: highway, urban road, country road, and provincial road, and the traffic flow can include three states: congested, moderate, and unobstructed. The above road types and traffic flows can be combined to obtain 12 driving conditions. When the vehicle travels on a specified driving condition each time, it can record the driving style based on the driver's behavior data, store the most frequently occurring driving style corresponding to the specified driving condition, and when the vehicle travels on the specified driving condition subsequently, it can directly read the driving style corresponding to the specified driving condition.

[0063] S304, process the target driving style, the environmental data of the vehicle, and the status information of the driver through a pre-trained assisted driving model to obtain a target control signal.

[0064] S305, obtain the actual control signal output by the intelligent driving domain controller of the vehicle based on the target intelligent driving algorithm.

[0065] S306, when the difference between the actual control signal and the target control signal is greater than a preset difference, determine the target parameter corresponding to the target driving style.

[0066] The preset difference is set according to experiments or experience and is not limited here. The fact that the difference between the actual control signal and the target control signal is greater than the preset difference indicates that the current parameter of the target intelligent driving algorithm does not match the driver's driving style, so it needs to be adjusted.

[0067] In some embodiments, the vehicle stores the mapping relationship between different driving styles and different parameters of the target intelligent driving algorithm. By querying the above mapping relationship, the target parameter corresponding to the target driving style can be obtained.

[0068] S307, adjust the current parameter of the target intelligent driving algorithm to the target parameter.

[0069] When the current parameter of the target intelligent driving algorithm is replaced with the target parameter, the parameter adjustment can be achieved.

[0070] In other possible implementation manners, the vehicle adjusts the current parameter of the target intelligent driving algorithm in multiple times, and compares the control signal output based on the parameter after each adjustment with the target control signal. If the difference between the two is less than the preset difference, the parameter after this adjustment is also the target parameter.

[0071] S308, control the vehicle to travel according to the target parameter.

[0072] In some embodiments, after the vehicle adjusts the parameters, it will also collect the driver's feedback information. If the driver's feedback information indicates satisfaction with the parameter adjustment result, the driving style recognition model and the assisted driving model will be iterated based on the environmental data, status information, driving style, etc. collected during the current driving process to further optimize the driving style recognition model and the assisted driving model, making the output results of both more accurate. Among them, the driver's feedback information includes the driver's emotion and the driver's behavior data of the vehicle. If at least one of the driver's emotion is a negative emotion or the driver intervenes in the intelligent assisted driving process occurs, the driver's feedback information indicates dissatisfaction with the parameter adjustment result; otherwise, the driver's feedback information indicates satisfaction with the parameter adjustment result. The driver's emotion can be determined by analyzing the face image collected by the image acquisition device installed in the vehicle and the voice information collected by the voice acquisition device.

[0073] In summary, the technical solution provided by the embodiments of the present application, during the driving process of the vehicle, determines the driver's driving style based on the real-time collected environmental data of the vehicle and the driver's status information, and then determines the target control signal based on the above environmental data, status information, and the determined driving style, and adjusts at least one parameter of the intelligent driving algorithm according to the target control signal and the actual control signal output by the intelligent driving algorithm, and then controls the vehicle to drive according to the adjusted parameters. In this way, the intelligent driving domain controller can adapt the control of the vehicle to the driving styles of the driver in different scenarios and different states, improving the driver's personalized assisted driving experience.

[0074] The following combines Figure 4 to elaborate on the training process of the driving style recognition model.

[0075] S401, obtain the first training sample set.

[0076] The first training sample set includes multiple groups of first training data. Each group of first training data in the multiple groups of first training data includes the environmental data of the vehicle, the driver's status information, and the driver's driving style. The quantity of the first training data can be determined according to the accuracy requirement of the driving style recognition model. The higher the accuracy requirement of the driving style recognition model, the greater the quantity of the first training data should be.

[0077] Optionally, S401 is specifically implemented as:

[0078] S4011, obtain multiple groups of historical driving data.

[0079] Each group of historical driving data in the multiple groups of historical driving data includes the environmental data of the vehicle, the driver's status information, and the driver's behavior data collected during the historical driving process.

[0080] Among them, the environmental data of the vehicle is collected through an environmental perception system during historical driving. The driver's status information is collected through a status collection device during historical driving. The driver's behavior data is used to record the operations performed by the driver on the actuators of the vehicle during historical driving, including but not limited to: the operation information of the drive pedal (such as the start time when the drive pedal is depressed, the duration, and the stroke of the drive pedal), the operation information of the brake pedal, the operation information of the steering wheel (such as the rotation direction and angle of the steering wheel), the operation information of the gear shift lever (such as the shifted gear, etc.).

[0081] During each driving process of the vehicle, the environmental data of the vehicle, the driver's status information, and the driver's behavior data are collected as a set of historical driving data and reported to the server.

[0082] S4012, Select K groups of data from multiple groups of historical driving data as the cluster centers of K clusters respectively.

[0083] K is an integer greater than 1. K is the number of clusters, which can be custom-set by technicians. In the embodiments of the present application, only K = 5 is taken as an example for illustration. The 5 clusters are also 5 driving styles, namely normal, aggressive, cautious, comfortable, and agile.

[0084] S4013, For the other groups of historical driving data except the K groups of data in the multiple groups of historical driving data, calculate the distances between the other groups of historical driving data and the K groups of data respectively to determine the clusters to which the other groups of historical driving data belong.

[0085] For any group of data in the other groups of data, the server calculates the Euclidean distances between this group of data and the K cluster centers respectively, and determines the cluster to which this group of data belongs as the cluster to which the target cluster center with the smallest Euclidean distance from this group of data belongs.

[0086] S4014, Update the cluster centers of the K clusters.

[0087] When the clusters to which the above other groups of data belong are all determined, update the cluster centers of each cluster based on the historical driving data included in each cluster.

[0088] S4015, Repeat the steps of calculating the clusters between the other groups of historical driving data and the K groups of data respectively and updating the cluster centers of the K clusters until the iteration stops after meeting the second stop iteration condition, and obtain the historical driving data sets corresponding to the K clusters respectively.

[0089] The second stop iteration condition means that the cluster centers of the K clusters no longer change, or the change value is less than a preset difference.

[0090] S4016. Based on the historical driving data sets corresponding to K clusters, store the environmental data of the vehicle, the status information of the driver, and the corresponding cluster in each set of historical driving data respectively to obtain the first training sample data set.

[0091] After the clustering is completed, the server stores the environmental data of the vehicle, the status information of the driver, and the corresponding cluster in each set of the above historical driving data as a set of first training data.

[0092] S402. For each set of the first training data in multiple sets of the first training data, input each set of the first training data into the first neural network model to obtain the output driving style.

[0093] The first neural network model can be

[0094] S503. Calculate the first loss function based on the output driving style and the above driving style.

[0095] The first loss function is used to measure the relative error between the output driving style and the above driving style.

[0096] S504. Update the parameters of the first neural network model through the first optimization algorithm to minimize the first loss function.

[0097] Optionally, the first optimization algorithm is the backpropagation algorithm.

[0098] S505. Repeat the steps of calculating the first loss function and updating the parameters of the first neural network model until the first stop iteration condition is satisfied and then stop the iterative training.

[0099] The first stop iteration condition can be that the first loss function is less than the first preset value, and the first preset value is set according to experiments or experience. The embodiments of the present application do not limit this. The stop iteration condition can also be that the number of iterations is greater than the first preset number of times, and the first preset number of times is set according to experiments or experience.

[0100] The following combines Figure 5 to elaborate on the training process of the assisted driving model. The training process of the assisted driving model includes:

[0101] S501. Obtain the second training sample set.

[0102] The second training sample set includes multiple sets of second training data, and the number of the second training data is determined according to the accuracy requirement of the assisted driving model. The higher the accuracy requirement of the assisted driving model, the more the number of the second training data.

[0103] Each group of second training data includes the environmental data of the vehicle, the status information of the driver, the driving style of the driver, and the behavior data of the driver. The determination process of the driving style of the driver can refer to the above S4041 - S4045, which will not be elaborated here. During each driving process of the vehicle, the above-mentioned environmental data of the vehicle, the status information of the driver, and the behavior data of the driver are collected and reported to the server. The server determines the driving style of the driver based on the received information, and then stores the received information corresponding to the determined driving style to obtain a group of second training data.

[0104] S502. For each group of second training data in multiple groups of second training data, input each group of second training data into the second neural network model to obtain output behavior data.

[0105] S503. Calculate the second loss function based on the output behavior data and the behavior data in the second training data.

[0106] The second loss function is used to measure the relative error between the output behavior data and the behavior data in the second training data.

[0107] S504. Update the parameters of the second neural network model through the second optimization algorithm to minimize the second loss function.

[0108] Optionally, the second optimization algorithm is the backpropagation algorithm.

[0109] S505. Repeat the steps of calculating the second loss function and updating the parameters of the second neural network model until the second stop iteration condition is met, and then stop the iterative training to obtain the assisted driving model.

[0110] The second stop iteration condition can be that the second loss function is less than the second preset value, and the second preset value is set according to experiments or experience, which is not limited in the embodiments of the present application. The second stop iteration condition can also be that the number of iterations is greater than the second preset number, and the second preset number is set according to experiments or experience.

[0111] Please refer to Figure 6 , which shows the block diagram of the device for controlling a vehicle shown in an embodiment of the present application. The device includes: a style determination module 610, a first acquisition module 620, a second acquisition module 630, a parameter adjustment module 640, and a vehicle control module 650.

[0112] The style determination module 610 is used to determine the target driving style of the driver of the vehicle during the driving process of the vehicle.

[0113] The first acquisition module 620 is used to process the target driving style, the environmental data of the vehicle, and the status information of the driver through a pre-trained assisted driving model to obtain a target control signal.

[0114] A second acquisition module 630, configured to acquire an actual control signal output by an intelligent driving domain controller of a vehicle based on a target intelligent driving algorithm.

[0115] A parameter adjustment module 640, configured to adjust at least one parameter of the target intelligent driving algorithm based on the actual control signal and the target control signal.

[0116] A vehicle control module 650, configured to control the vehicle to travel according to the adjusted parameters.

[0117] In some embodiments, the parameter adjustment module 640 is configured to determine target parameters corresponding to a target driving style when a difference between the actual control signal and the target control signal is greater than a preset difference; adjust the current parameters of the target intelligent driving algorithm to the target parameters; and control the vehicle to travel according to the target parameters.

[0118] In some embodiments, a style determination module 610 is configured to collect environmental data; acquire status information; and process the environmental data and the status information through a pre-trained driving style recognition model to obtain a target driving style.

[0119] In some embodiments, the training process of the driving style recognition model includes: acquiring a first training sample data set, where the first training data set includes multiple groups of first training data, and each group of first training data in the multiple groups of first training data includes environmental data of a vehicle, status information of a driver, and the driving style of the driver; for each group of first training data, inputting each group of first training data into a first neural network model to obtain an output driving style; calculating a first loss function based on the output driving style and the driving style in each group of first training data; updating model parameters in the first neural network model through a first optimization algorithm to minimize the first loss function; and repeating the steps of calculating the first loss function and updating the model parameters in the first neural network model until the first stop iteration condition is satisfied and then stopping the iterative training to obtain the driving style recognition model.

[0120] In some embodiments, obtaining the first training sample data set includes: obtaining multiple sets of historical driving data, where each set of historical driving data in the multiple sets of historical driving data includes the environmental data of the vehicle, the driver's status information, and the driver's behavior data collected during the historical driving process; selecting K sets of data from the multiple sets of historical driving data as the clustering centers of K clusters respectively, where K is an integer greater than 1; for the other sets of historical driving data except the K sets of data in the multiple sets of historical driving data, calculating the clustering between the other sets of historical driving data and the K sets of data respectively to determine the clusters to which the other sets of historical driving data belong; updating the clustering centers of the K clusters; repeating the steps of calculating the clustering between the other sets of historical driving data and the K sets of data respectively and updating the clustering centers of the K clusters until the iteration stops after meeting the second stop iteration condition, obtaining the historical driving data sets corresponding to the K clusters respectively; based on the historical driving data sets corresponding to the K clusters respectively, storing the environmental data of the vehicle, the driver's status information, and the corresponding clusters in each set of historical driving data correspondingly to obtain the first training sample data set.

[0121] In some embodiments, the style determination module 610 is configured to collect the environmental data of the vehicle; determine the driving road condition of the vehicle based on the environmental data of the vehicle; obtain the driving style corresponding to the driving road condition of the vehicle as the target driving style.

[0122] In some embodiments, the training process of the assisted driving model includes: obtaining a second training sample data set, where the second training data set includes multiple sets of second training data, and each set of second training data in the multiple sets of second training data includes the environmental data of the vehicle, the driver's status information, the driver's driving style, and the driver's behavior data; for each set of second training data, inputting each set of second training data into the second neural network model to obtain the output behavior data; calculating the second loss function based on the output behavior data and the behavior data in each set of second training data; updating the model parameters in the second neural network model through the second optimization algorithm to minimize the second loss function; repeating the steps of calculating the second loss function and updating the model parameters in the second neural network model until the iteration training stops after meeting the second stop iteration condition, obtaining the assisted driving model.

[0123] In summary, the technical solution provided by the embodiments of the present application, during the driving process of the vehicle, first determines the driver's driving style, and then determines the target control signal based on the real-time collected environmental data of the vehicle, the driver's status information, and the determined driving style, and adjusts at least one parameter of the intelligent driving algorithm according to the target control signal and the actual control signal output by the intelligent driving algorithm, and then controls the vehicle to drive according to the adjusted parameters. In this way, the intelligent driving domain controller can adapt to the driving styles of the driver in different scenarios and different states, improving the driver's personalized assisted driving experience.

[0124] Please refer to Figure 7 which shows that an embodiment of the present application further provides a vehicle 700, which includes: one or more processors 710, a memory 720, and one or more applications. Among them, the one or more applications are stored in the memory 720 and configured to be executed by the one or more processors 710, and the one or more applications are configured to execute the methods described in the above embodiments.

[0125] The processor 710 may include one or more processing cores. The processor 710 connects various parts within the entire battery management system through various interfaces and lines, and executes various functions of the battery management system and processes data by running or executing instructions, programs, code sets, or instruction sets stored in the memory 720, and by calling data stored in the memory 720. Optionally, the processor 710 may be implemented in at least one hardware form of digital signal processing (DSP), field-programmable gate array (FPGA), or programmable logic array (PLA). The processor 710 may integrate one or a combination of several of a central processing unit 710 (CPU), a graphics processing unit 710 (GPU), and a modem. Among them, the CPU mainly processes the operating system, user interface, and applications, etc.; the GPU is responsible for rendering and drawing display content; the modem is used to process wireless communication. It can be understood that the above modem may not be integrated into the processor 710 and may be implemented separately through a communication vehicle.

[0126] The memory 720 may include a random access memory 720 (RAM), and may also include a read-only memory 720 (ROM). The memory 720 can be used to store instructions, programs, codes, code sets, or instruction sets. The memory 720 may include a program storage area and a data storage area. Among them, the program storage area may store instructions for implementing the operating system, instructions for implementing at least one function (such as touch function, sound playback function, image playback function, etc.), instructions for implementing the following various method embodiments, etc. The data storage area may also store data created during the use of the vehicle map (such as phone book, audio and video data, chat record data, etc.).

[0127] The embodiments of the present application also provide a computer-readable storage medium, in which computer program instructions are stored, and the computer program instructions can be called by a processor to execute the methods described in the above embodiments.

[0128] The computer-readable storage medium can be an electronic memory such as flash memory, EEPROM (electrically erasable programmable read-only memory), EPROM, hard disk, or ROM. Optionally, the computer-readable storage medium includes a non-transitory computer-readable storage medium. The computer-readable storage medium has a storage space for computer program instructions for executing any method steps in the above methods. These computer program instructions can be read from or written into one or more computer program products. The computer program instructions can be compressed in a suitable form.

[0129] The above are only the preferred embodiments of the present application and do not impose any form of limitation on the present application. Although the present application has been disclosed above with the preferred embodiments, it is not intended to limit the present application. Any person skilled in the art can make some changes or modifications to equivalent embodiments with equivalent changes within the scope of the technical solutions of the present application. However, as long as it does not deviate from the content of the technical solutions of the present application, any brief modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present application still fall within the scope of the technical solutions of the present application.

Claims

1. A method for controlling a vehicle, characterized in that, The method includes: During the driving process of the vehicle, determining the target driving style of the driver of the vehicle; Processing the target driving style, the environmental data of the vehicle collected in real time, and the status information of the driver collected in real time through a pre-trained assisted driving model to obtain a target control signal; Obtaining the actual control signal output by the intelligent driving domain controller of the vehicle based on the target intelligent driving algorithm; Adjusting at least one parameter of the target intelligent driving algorithm based on the actual control signal and the target control signal; Controlling the vehicle to drive according to the adjusted parameter.

2. The method according to claim 1, characterized in that The adjusting at least one parameter of the target intelligent driving algorithm of the vehicle based on the actual control signal and the target control signal includes: When the difference between the actual control signal and the target control signal is greater than a preset difference, determining the target parameter corresponding to the target driving style; Adjusting the current parameter of the target intelligent driving algorithm to the target parameter; The controlling the vehicle to drive according to the adjusted intelligent driving parameter includes: Controlling the vehicle to drive according to the target parameter.

3. The method according to claim 1, characterized in that, The determining the target driving style of the driver of the vehicle during the driving process of the vehicle includes: Collecting the environmental data; Obtaining the status information; Processing the environmental data and the status information through a pre-trained driving style recognition model to obtain the target driving style.

4. The method according to claim 3, wherein The training process of the driving style recognition model includes: Obtaining a first training sample data set, the first training data set includes multiple groups of first training data, and each group of first training data in the multiple groups of first training data includes the environmental data of the vehicle, the status information of the driver, and the driving style of the driver; For each group of first training data, inputting each group of first training data into a first neural network model to obtain an output driving style; Calculating a first loss function based on the output driving style and the driving style in each group of first training data; Updating the model parameters in the first neural network model through a first optimization algorithm to minimize the first loss function; Repeating the steps of calculating the first loss function and updating the model parameters in the first neural network model until the first stop iteration condition is met and then stopping the iterative training to obtain the driving style recognition model.

5. The method according to claim 4, wherein The obtaining the first training sample data set includes: Obtaining multiple groups of historical driving data, and each group of historical driving data in the multiple groups of historical driving data includes the environmental data of the vehicle, the status information of the driver, and the behavior data of the driver collected during the historical driving process; Selecting K groups of data from the multiple groups of historical driving data as the clustering centers of K clusters respectively, where K is an integer greater than 1; For the other groups of historical driving data in the multiple groups of historical driving data except the K groups of data, respectively calculating the clustering between the other groups of historical driving data and the K groups of data to determine the clusters to which the other groups of historical driving data belong; Updating the clustering centers of the K clusters; Repeat the steps of separately calculating the clustering between the historical driving data of the other groups and the K groups of data and updating the cluster centers of the K clusters until the iteration stops after meeting the second stop iteration condition, and obtain the historical driving data sets corresponding to the K clusters respectively; Based on the historical driving data sets corresponding to the K clusters respectively, store the environmental data of the vehicle, the status information of the driver, and the corresponding clusters in each group of historical driving data correspondingly, and obtain the first training sample data set.

6. The method according to claim 1, characterized in that Determining the target driving style of the driver of the vehicle during the driving of the vehicle includes: Collect the environmental data of the vehicle; Determine the driving road conditions of the vehicle based on the environmental data of the vehicle; Obtain the driving style corresponding to the driving road conditions of the vehicle as the target driving style.

7. The method according to any one of claims 1 to 6, characterized in that, The training process of the assisted driving model includes: Obtain a second training sample data set, where the second training data set includes multiple groups of second training data, and each group of second training data in the multiple groups of second training data includes the environmental data of the vehicle, the status information of the driver, the driving style of the driver, and the behavior data of the driver; For each group of second training data, input the each group of second training data into a second neural network model to obtain output behavior data; Calculate a second loss function based on the output behavior data and the behavior data in each group of second training data; Update the model parameters in the second neural network model through a second optimization algorithm to minimize the second loss function; Repeat the steps of calculating the second loss function and updating the model parameters in the second neural network model until the iteration training stops after meeting the second stop iteration condition, and obtain the assisted driving model.

8. A device for controlling a vehicle, characterized in that, The device includes: A style determination module, configured to determine the target driving style of the driver of the vehicle during the driving of the vehicle; A first acquisition module, configured to process the target driving style, the environmental data of the vehicle collected in real time, and the status information of the driver collected in real time through a pre-trained assisted driving model to obtain a target control signal; A second acquisition module, configured to obtain an actual control signal output by the intelligent driving domain controller of the vehicle based on a target intelligent driving algorithm; A parameter adjustment module, configured to adjust at least one parameter of the target intelligent driving algorithm based on the actual control signal and the target control signal; A vehicle control module, configured to control the vehicle to drive according to the adjusted parameters.

9. A vehicle, characterized in that, Includes: A memory; One or more processors, coupled to the memory; One or more applications, where the one or more applications are stored in the memory and are configured to be executed by the one or more processors, and the one or more applications are configured to execute the method according to any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, Computer program instructions are stored in the computer-readable storage medium, and the computer program instructions can be called by a processor to execute the method according to any one of claims 1-7.

Images