Vehicle control method and device, vehicle, storage medium and program product
By determining the target control parameters based on the user's vehicle control habits, the problem that the existing autonomous driving system cannot meet personalized driving habits is solved, and the matching degree of autonomous driving experience and driving habits is improved.
Patent Information
- Application Number
- CN202411367755.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-27
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2044-09-27
AI Technical Summary
The existing autonomous driving system cannot meet the personalized driving habits of all users, resulting in poor autonomous driving experience.
By determining the target control parameters based on the user's vehicle control habits, processing historical data using the target action model, generating control parameters that conform to the user's driving habits, thereby controlling the vehicle.
It improves the user's autonomous driving experience, reduces manual intervention, and is more in line with the user's driving habits.
Smart Images

Figure CN118928373A_ABST
Abstract
Description
Technical Field
[0001] The present disclosure relates to the field of vehicle control technology, and in particular to a vehicle control method, device, vehicle, storage medium and program product. Background Art
[0002] As an important direction for the future development of the automotive industry, autonomous driving technology is accelerating its development around the world. With the development of autonomous driving technology, users have higher expectations for the autonomous driving vehicle experience.
[0003] In the related art, the automatic driving system adopts universal driving parameters, which cannot meet the personalized driving habits of all users. For example, for the control of adaptive cruise, the vehicle is controlled by multiple fixed following distances. Summary of the invention
[0004] In order to overcome the problems existing in the related art, the present disclosure provides a vehicle control method, device, vehicle, storage medium and program product to determine the target control parameters based on the user's vehicle control habits, so as to better simulate the user's driving habits and improve the user's autonomous driving experience.
[0005] According to a first aspect of an embodiment of the present disclosure, there is provided a vehicle control method, comprising: Determine the target action corresponding to the target driving scenario; The target data is processed by a target action model corresponding to the target action to obtain a target control parameter, wherein the target data is data related to executing the target action, and the target action model is obtained by training a basic model based on a plurality of sample data, and the sample data is historical data related to the target action during the historical driving process of the target user; The vehicle is controlled according to the target control parameter.
[0006] Optionally, the target action model is obtained by the following steps: Acquire a plurality of sample data related to the target action during the target user's historical driving process, wherein the sample data includes sample association parameters and actual control parameters; Performing multiple rounds of iterative training on the basic model using sample association parameters in the multiple sample data; After each round of training, the prediction control parameters corresponding to this round of training are obtained; The prediction loss corresponding to this round of training is obtained through the prediction control parameters corresponding to this round of training and the actual control parameters in the sample data corresponding to this round of training; Optimizing the basic model through the prediction loss corresponding to this round of training; When the basic model meets the preset conditions, the training is stopped to obtain the target action model.
[0007] Optionally, the target action includes a turn signal operation, the target action model includes a turn signal model, and the target data includes a target road condition and a target driving scene; The target data is processed by the target action model corresponding to the target action to obtain the target control parameter, including: The target road condition and the target driving scenario are processed by the turn signal model to obtain turn signal control parameters corresponding to the turn signal operation. The turn signal model is obtained by training a basic turn signal model based on multiple sample turning data. The sample turning data are sample road conditions, sample driving scenarios and actual turn signal parameters related to the turn signal operation during the historical driving process of the target user.
[0008] Optionally, the target action includes pedal operation, the target action model includes a pedal model, and the target data includes a target speed and a target traffic condition; The target data is processed by the target action model corresponding to the target action to obtain the target control parameter, including: The target speed and the target traffic condition are processed by the pedal model to obtain pedal control parameters corresponding to the pedal operation. The pedal model is obtained by training a basic pedal model based on multiple sample pedal data. The sample pedal data are sample speeds, sample traffic conditions and actual pedal parameters related to the pedal operation during the historical driving process of the target user.
[0009] Optionally, the target action includes a speed adjustment operation, the target action model includes a speed adjustment model, and the target data includes a target road section and a target traffic condition; The target data is processed by the target action model corresponding to the target action to obtain the target control parameter, including: The target road section and the target traffic condition are processed by the speed adjustment model to obtain a speed control parameter corresponding to the speed adjustment operation. The speed adjustment model is obtained by training a basic speed model based on a plurality of sample speed data, and the sample speed data are sample road sections, sample traffic conditions and actual speed parameters related to the speed adjustment operation during the historical driving process of the target user.
[0010] Optionally, the target action includes a following distance adjustment operation, the target action model includes a following distance adjustment model, and the target data includes a target speed and a target traffic condition; The target data is processed by the target action model corresponding to the target action to obtain the target control parameter, including: The target speed and the target traffic condition are processed by the following distance adjustment model to obtain a following distance control parameter corresponding to the following distance adjustment operation. The following distance adjustment model is obtained by training a basic distance model based on a plurality of sample distance data, and the sample distance data are sample speeds, sample traffic conditions and actual distance parameters related to the following distance adjustment operation during the historical driving process of the target user.
[0011] Optionally, the target action includes an obstacle avoidance operation, the target action model includes an obstacle avoidance model, and the target data includes a target obstacle and a target vehicle state; The target data is processed by the target action model corresponding to the target action to obtain the target control parameter, including: The target obstacle and the target vehicle state are processed by the obstacle avoidance model to obtain an obstacle avoidance trajectory corresponding to the obstacle avoidance operation. The obstacle avoidance model is obtained by training a basic obstacle avoidance model based on multiple sample obstacle avoidance data. The sample obstacle avoidance data are sample obstacles, sample vehicle states and actual obstacle avoidance trajectories related to the obstacle avoidance operation during the target user's historical driving process.
[0012] Optionally, determining a target action corresponding to the target driving scenario includes: Determine the target control strategy according to vehicle driving parameters and environmental parameters; Determining the target driving scenario according to the target control strategy; Based on the association relationship between the driving scene and the execution action, the target action corresponding to the target driving scene is determined, and one target driving scene corresponds to at least one target action.
[0013] Optionally, determining the target control strategy according to the vehicle driving parameters and the environmental parameters includes: The target control strategy is obtained by processing the vehicle driving parameters and the environmental parameters through a target strategy model, and the target strategy model is obtained by training a basic strategy model based on multiple sample strategy data, and the sample strategy data is historical data related to the vehicle's intervention actions with the target user in the automatic driving mode.
[0014] According to a second aspect of an embodiment of the present disclosure, there is provided a vehicle control device, comprising: A determination module, configured to determine a target action corresponding to a target driving scenario; A first obtaining module is configured to process target data through a target action model corresponding to the target action to obtain a target control parameter, wherein the target data is data related to executing the target action, and the target action model is obtained by training a basic model based on a plurality of sample data, and the sample data is historical data related to the target action during the historical driving process of the target user; The control module is configured to control the vehicle according to the target control parameter.
[0015] According to a third aspect of an embodiment of the present disclosure, a vehicle is provided, comprising: processor; a memory for storing processor-executable instructions; Among them, the processor is configured to implement the steps of the vehicle control method provided by the first aspect of the present disclosure when executed.
[0016] According to a fourth aspect of an embodiment of the present disclosure, a computer-readable storage medium is provided, on which computer program instructions are stored. When the computer program instructions are executed by a processor, the steps of the vehicle control method provided in the first aspect of the present disclosure are implemented.
[0017] According to a fifth aspect of an embodiment of the present disclosure, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the steps of the vehicle control method provided in the first aspect of the present disclosure.
[0018] The technical solution provided by the embodiments of the present disclosure may have the following beneficial effects: First, determine the target action corresponding to the target driving scenario, and then process the target data through the target action model corresponding to the target action to obtain the target control parameters. The target data is data related to executing the target action. The target action model is obtained by training a basic model based on multiple sample data. The sample data is historical data related to the target action during the target user's historical driving process. According to the target control parameters, the vehicle is controlled to obtain a vehicle control method that is more in line with the user's driving habits.
[0019] Among them, the basic model is trained by using the historical data related to the target action during the target user's historical driving process as sample data, so that a target action model corresponding to the target action and in line with the user's driving habits can be obtained. Then, when the vehicle needs to perform the corresponding target action, the target action model is used to process the data related to the execution of the target action to obtain target control parameters that meet the user's driving habits, and the vehicle is controlled by the target control parameters, thereby better simulating the user's driving habits, minimizing the user's manual intervention, and improving the user's autonomous driving experience.
[0020] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS
[0021] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the present disclosure.
[0022] Figure 1 The diagram is a schematic diagram of a scenario of a vehicle control method according to an exemplary embodiment.
[0023] Figure 2 The present invention is a flow chart of a vehicle control method according to an exemplary embodiment.
[0024] Figure 3 is a block diagram of a vehicle control device according to an exemplary embodiment.
[0025] Figure 4 is a block diagram of a vehicle according to an exemplary embodiment. DETAILED DESCRIPTION
[0026] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. Instead, they are merely examples of devices and methods consistent with some aspects of the present disclosure as detailed in the appended claims.
[0027] It should be noted that all actions of acquiring signals, information or data in the present disclosure are carried out in compliance with the relevant data protection laws and policies of the country where the device is located and with the authorization given by the owner of the corresponding device.
[0028] In the related art, the automatic driving system adopts universal driving parameters, which cannot meet the personalized driving habits of all users. For example, for the control of adaptive cruise, the vehicle is controlled by multiple fixed following distances.
[0029] In response to the above technical problems, the embodiments of the present disclosure provide a vehicle control method, device, vehicle, storage medium and program product. By using historical data related to the target action of the target user in the historical driving process as sample data to train the basic model, a target action model corresponding to the target action and conforming to the user's driving habits can be obtained. Then, when the vehicle needs to perform the corresponding target action, the target action model is used to process the data related to the execution of the target action to obtain target control parameters that conform to the user's driving habits, and the vehicle is controlled by the target control parameters, thereby better simulating the user's driving habits, minimizing the user's manual intervention, and improving the user's autonomous driving experience.
[0030] Figure 1 is a schematic diagram of a vehicle control method according to an exemplary embodiment. Figure 1 As shown, this vehicle, i.e. the vehicle driven by the target user, may encounter driving scenarios such as lane changing, overtaking and lane changing, lane keeping, turning, and obstacle avoidance when driving on the road.
[0031] Figure 2 is a flow chart of a vehicle control method according to an exemplary embodiment. Figure 2 As shown, the following steps may be included.
[0032] In step S201, a target action corresponding to a target driving scene is determined.
[0033] During the driving process of the vehicle, various driving scenarios will be encountered, such as lane change, lane change, lane keeping, turning, obstacle avoidance, etc. Each driving scenario requires at least one action to complete. The association relationship between different driving scenarios and execution actions can be established in advance, and then when the target driving scenario is determined, the target action corresponding to the target driving scenario can be determined through the association relationship.
[0034] In step S202, the target data is processed by a target action model corresponding to the target action to obtain target control parameters. The target data is data related to executing the target action. The target action model is obtained by training a basic model based on multiple sample data. The sample data is historical data related to the target action during the target user's historical driving process.
[0035] In this embodiment, the target action model corresponding to the target action and in line with the user's driving habits can be obtained by training the basic model with historical data related to the target action during the target user's historical driving process as sample data. Then, when the vehicle needs to perform the corresponding target action, the target action model is used to process the data related to the execution of the target action to obtain the target control parameters that meet the user's driving habits, thereby better simulating the user's driving habits. Among them, the basic model can be MILE (Model-Based Imitation Learning for Urban Driving, imitation learning framework).
[0036] In step S203, the vehicle is controlled according to the target control parameters.
[0037] In this embodiment, the vehicle is controlled by target control parameters, so as to better simulate the user's driving habits, minimize the user's manual intervention, and improve the user's autonomous driving experience.
[0038] In a possible implementation, the target action model is obtained by the following steps: Acquire multiple sample data related to the target action during the target user's historical driving process, the sample data including sample association parameters and actual control parameters; perform multiple rounds of iterative training on the basic model through the sample association parameters in the multiple sample data; obtain the prediction control parameters corresponding to this round of training after each round of training; obtain the prediction loss corresponding to this round of training through the prediction control parameters corresponding to this round of training and the actual control parameters in the sample data corresponding to this round of training; optimize the basic model through the prediction loss corresponding to this round of training; when the basic model meets the preset conditions, stop the training and obtain the target action model.
[0039] In this embodiment, the basic model can be iteratively trained for multiple rounds using multiple sample data so as to continuously optimize the basic model so that the basic model can meet preset conditions and the output results can meet the user's driving habits, thereby obtaining a target action model.
[0040] Specifically, the basic model can be trained by the sample association parameters in multiple sample data. In each round of training, the multiple sample association parameters corresponding to the round of training are input into the basic model, and the prediction control parameters corresponding to the current round of training can be output through the basic model. Then, the prediction loss corresponding to the current round of training is obtained by the prediction control parameters corresponding to the current round of training and the actual control parameters in the sample data corresponding to the current round of training. The distance between the prediction control parameters corresponding to the current round of training and the actual control parameters in the sample data corresponding to the current round of training can be calculated, and the distance can be used as the corresponding prediction loss. And the parameters in the basic model can be optimized by the prediction loss based on the gradient descent algorithm. Each round of training can be optimized once until the basic model meets the preset conditions. The preset conditions can be that the number of training rounds reaches the preset number of rounds, or the basic model converges. When the basic model meets the preset conditions, the training can be stopped, and the basic model after the last optimization is determined as the target action model. It is also possible to perform target recognition on users. Considering that different users may have different driving habits, the corresponding relevant historical data for the same target action may also be different. For the vehicle driven by the target user, it is necessary to use the historical data related to the target action during the target user's historical driving process as sample data to train the corresponding target action model.
[0041] By training the basic model with relevant data when the user performs the target action during manual driving of the vehicle, it is possible to train a target action model corresponding to the target action that conforms to the user's driving habits, thereby being able to obtain target control parameters that conform to the user's driving habits based on the target action model to control the vehicle.
[0042] In a possible implementation, the target action includes a turn signal operation, the target action model includes a turn signal model, and the target data includes a target road condition and a target driving scene; The target data is processed by the target action model corresponding to the target action to obtain the target control parameters, including: The target road condition and target driving scenario are processed through the turn signal model to obtain the turn signal control parameters corresponding to the turn signal operation. The turn signal model is obtained by training the basic turn signal model based on multiple sample turning data. The sample turning data are sample road conditions, sample driving scenarios and actual turn signal parameters related to the turn signal operation of the target user in the historical driving process.
[0043] In this embodiment, the target action may include a turn signal operation, which can be used in scenes of turning, changing lanes, avoiding obstacles, and changing lanes when overtaking. The basic turn signal model can be trained based on historical data related to the turn signal operation of the target user during historical driving. The specific training method can refer to the training method for obtaining the target action model mentioned above, which will not be repeated here. Among them, the sample road condition and the sample driving scene are sample association parameters, and the actual turn signal parameter is the actual control parameter. The target road condition can be one of multiple road surface grades.
[0044] The actual turn signal parameters may be parameters related to the turn signal operation of the target user during driving, and may include the operation L of using the turn signal, such as turning left or turning right, the operation timing T, the frequency F, and the turn signal duration D. The user's turn signal operation can be represented by the function g(L, T, F, D).
[0045] The turn signal operation of the autonomous driving system simulating the user's operation can be expressed by the following function: L_auto = g(L_user)×K Among them, K is an adjustment coefficient, different target data corresponds to different K values, L_user represents the user's operating habits, that is, (L, T, F, D), and L_auto is the operation simulated by the system, that is, the turn signal control parameter.
[0046] By processing the target road condition and the target driving scenario through the turn signal model, the turn signal control parameters corresponding to the turn signal operation can be obtained. The turn signal control parameters may include the operation of using the turn signal, the operation timing, the frequency and the turn signal duration.
[0047] In a possible implementation, the target action includes pedal operation, the target action model includes a pedal model, and the target data includes a target speed and a target traffic condition; The target data is processed by the target action model corresponding to the target action to obtain the target control parameters, including: The target speed and target traffic conditions are processed through the pedal model to obtain pedal control parameters corresponding to the pedal operation. The pedal model is obtained by training a basic pedal model based on multiple sample pedal data. The sample pedal data are sample speeds, sample traffic conditions and actual pedal parameters related to the pedal operation of the target user during historical driving.
[0048] In this embodiment, the target action may include pedal operation, which can be used in scenes such as turning, changing lanes, avoiding obstacles, keeping lanes, and overtaking and changing lanes. The basic pedal model can be trained based on historical data related to pedal operation during the target user's historical driving process. The specific training method can refer to the training method for obtaining the target action model mentioned above, which will not be repeated here. Among them, the sample speed and the sample traffic condition are the sample association parameters, and the actual pedal parameter is the actual control parameter.
[0049] The pedals include the accelerator pedal and the brake pedal. The actual pedal parameters can be the operation details of the accelerator and brake pedals when the target user accelerates or decelerates, such as pedal depth, operation smoothness, and reaction time. The target speed is the speed when the pedal operation is performed, and the target traffic condition can be one of a red light or a deceleration of the vehicle ahead.
[0050] The accelerator pedal depth is P_throttle, the brake pedal depth is P_brake, the smoothness of the operation is S, and the reaction time is R. The user's pedal operation can be described using the function h(P_throttle, P_brake, S, R).
[0051] The vehicle's automatic driving system simulates the user's pedal operation and can be expressed by the following function: P_auto = h(P_user)×β Among them, β is an adaptation coefficient, different target data corresponds to different β values, P_user represents the user's pedal operation habits, that is, (P_throttle, P_brake, S, R), and P_auto is the operation intensity simulated by the system, that is, the pedal control parameter.
[0052] By processing the target speed and the target traffic condition through the pedal model, the pedal control parameters corresponding to the pedal operation can be obtained. The pedal control parameters may include pedal depth, smoothness of operation and reaction time.
[0053] In a possible implementation, the target action includes a speed adjustment operation, the target action model includes a speed adjustment model, and the target data includes a target road section and a target traffic condition; The target data is processed by the target action model corresponding to the target action to obtain the target control parameters, including: The target road section and target traffic condition are processed through the speed adjustment model to obtain the speed control parameters corresponding to the speed adjustment operation. The speed adjustment model is obtained by training the basic speed model based on multiple sample speed data. The sample speed data are the sample road section, sample traffic condition and actual speed parameters related to the speed adjustment operation during the historical driving process of the target user.
[0054] In this embodiment, the target action may include a speed adjustment operation, which can be used in scenes such as turning, changing lanes, avoiding obstacles, keeping lanes, and overtaking and changing lanes. The speed adjustment can be completed in combination with the pedal operation. The basic speed model can be trained based on the historical data related to the speed adjustment operation during the historical driving process of the target user. The specific training method can refer to the training method for obtaining the target action model mentioned above, which will not be repeated here. Among them, the sample road section and the sample traffic condition are the sample association parameters, and the actual speed parameter is the actual control parameter.
[0055] The actual speed parameter may be the speed after the target user adjusts the speed in manual driving or automatic driving. The target road section is the road section where the user is when adjusting the speed, such as a city road or a highway. The target traffic condition may be the traffic condition where the user is when adjusting the speed, such as congestion or smooth traffic.
[0056] Through machine learning algorithms, the basic speed model is trained to learn the speed preferences of target users in different road sections and under different traffic conditions, and the speed adjustment model is obtained to obtain the speed that the target user may expect under specific road conditions, that is, the speed control parameter, such as the cruising speed during adaptive cruise control. This enables the vehicle to automatically adjust to the user's preferred speed without explicit instructions, improving the user's driving experience.
[0057] The speed preference of the target user is V_pref, and the speed selection function under different road sections L and different traffic conditions C is v(V_pref, L, C).
[0058] The vehicle's automatic driving system simulates the user's speed adjustment operation and can be expressed by the following function: V_cruise = v(V_pref, L, C) + δ Among them, δ is a deviation value adjusted according to historical data. Different target data corresponds to different δ values, which is used to fine-tune to be closer to the speed preference of the target user. V_cruise is the speed control parameter.
[0059] The target road section and the target traffic condition are processed by the speed adjustment model to obtain speed control parameters corresponding to the speed adjustment operation. The speed control parameters may include the expected speed.
[0060] In a possible implementation, the target action includes a following distance adjustment operation, the target action model includes a following distance adjustment model, and the target data includes a target speed and a target traffic condition; The target data is processed by the target action model corresponding to the target action to obtain the target control parameters, including: The target speed and target traffic condition are processed through the following distance adjustment model to obtain the following distance control parameters corresponding to the following distance adjustment operation. The following distance adjustment model is obtained by training a basic speed model based on multiple sample distance data. The sample distance data are the sample speed, sample traffic condition and actual distance parameters related to the following distance adjustment operation of the target user in the historical driving process.
[0061] In this embodiment, the target action may include a following distance adjustment operation, which can be used in any lane keeping scenario. The basic speed model can be trained based on historical data related to the following distance adjustment operation during the target user's historical driving process. The specific training method can refer to the training method for obtaining the target action model mentioned above, which will not be repeated here. Among them, the sample speed and the sample traffic condition are the sample association parameters, and the actual distance parameter is the actual control parameter.
[0062] The actual distance parameter may be the speed of the target user after adjusting the following distance in manual driving or automatic driving. The target speed is the speed of the user when adjusting the following distance, and the target traffic condition may be the traffic condition in which the user is when adjusting the following distance, such as congestion or smooth traffic.
[0063] Through machine learning algorithms, the basic distance model is trained to learn the target user's following distance preferences under different speeds and traffic conditions, and the following distance adjustment model is obtained to obtain the following distance that the target user may expect under specific road conditions, that is, the following distance control parameter, such as the following distance during adaptive cruise control. This allows the vehicle to automatically adjust to the user's preferred following distance without explicit instructions, improving the user's driving experience.
[0064] By processing the target speed and the target traffic condition through the following distance adjustment model, a following distance control parameter corresponding to the following distance adjustment operation can be obtained, and the following distance control parameter can be the expected following distance.
[0065] In a possible implementation, the target action includes an obstacle avoidance operation, the target action model includes an obstacle avoidance model, and the target data includes a target obstacle and a target vehicle state; The target data is processed by the target action model corresponding to the target action to obtain the target control parameters, including: The target obstacle and the target vehicle state are processed by the obstacle avoidance model to obtain the obstacle avoidance trajectory corresponding to the obstacle avoidance operation. The obstacle avoidance model is obtained by training the basic obstacle avoidance model based on multiple sample obstacle avoidance data. The sample obstacle avoidance data are sample obstacles, sample vehicle states and actual obstacle avoidance trajectories related to the obstacle avoidance operation of the target user in the historical driving process.
[0066] In this embodiment, the target action may include an obstacle avoidance operation, which can be used in the obstacle avoidance scenario. The basic obstacle avoidance model can be trained based on the historical data related to the obstacle avoidance operation during the historical driving process of the target user. The specific training method can refer to the training method for obtaining the target action model mentioned above, which will not be repeated here. Among them, the sample obstacle and the sample vehicle state are the sample association parameters, and the actual obstacle avoidance trajectory is the actual control parameter.
[0067] The actual obstacle avoidance trajectory can be the trajectory of the target user when performing obstacle avoidance in manual driving. The target obstacle is the obstacle that the user avoids when performing obstacle avoidance, such as road damage or obstacles on the road. The target vehicle state can be the state of the vehicle when the user performs obstacle avoidance, such as vehicle speed (v_vehicle), steering wheel angle (θ_steering), brake pressure (p_brake) and accelerator pedal position (p_accel).
[0068] Through machine learning algorithms, the basic obstacle avoidance model is trained to learn the target user's obstacle avoidance preferences under different obstacles and different vehicle conditions, and obtain an obstacle avoidance model in order to obtain the obstacle avoidance trajectory that the target user may expect under specific road conditions, such as the obstacle avoidance trajectory when encountering stationary stones or stationary vehicles on the road. This enables the vehicle to automatically adjust to the user's preferred obstacle avoidance trajectory without explicit instructions, improving the user's driving experience. During the training process, the ideal obstacle avoidance trajectory (T_ideal) and the time interval (t_reaction) from the time the user detects an obstacle to the time the user takes action can also be added to ensure the safety of autonomous driving, so that the user's obstacle avoidance habits can be met as much as possible while ensuring driving safety.
[0069] In a possible implementation, determining a target action corresponding to a target driving scenario includes: Determine the target control strategy based on vehicle driving parameters and environmental parameters; determine the target driving scene based on the target control strategy; determine the target action corresponding to the target driving scene based on the association between the driving scene and the execution action, and one target driving scene corresponds to at least one target action.
[0070] In this embodiment, the scenes of the vehicle during driving can be divided into multiple scenes, such as lane change, overtaking lane change, lane keeping, turning, obstacle avoidance and other driving scenes. Each driving scene requires at least one action to complete, and the association relationship between different driving scenes and execution actions can be established in advance. The vehicle driving parameters and environmental parameters can be obtained in advance to determine the target control strategy, wherein the vehicle driving parameters may include driving speed, acceleration, steering wheel angle, brake pressure and accelerator pedal position, etc., and the environmental parameters may include the lane where the vehicle is located, the road conditions, the distance from the front vehicle and the rear vehicle, the road section where the vehicle is located, and the position coordinates of the vehicle and the position coordinates of the vehicles around the vehicle, etc. A target driving scene corresponds to at least one target action. For example, when the target driving scene is lane change, overtaking lane change or turning, the corresponding target action may include turn signal operation, pedal operation and speed adjustment operation; when the target driving scene is lane keeping, the corresponding target action may include pedal operation, speed adjustment operation or following distance adjustment operation; when the target driving scene is obstacle avoidance, the corresponding target action may include obstacle avoidance operation.
[0071] After determining the target control strategy, the target driving scenario can be determined based on the current vehicle state and the vehicle state corresponding to the target control strategy, and then the target action corresponding to the target driving scenario can be determined based on the association between the driving scenario and the execution action.
[0072] In a possible implementation, determining a target control strategy according to vehicle driving parameters and environmental parameters includes: The vehicle driving parameters and environmental parameters are processed through the target strategy model to obtain the target control strategy. The target strategy model is obtained by training the basic strategy model based on multiple sample strategy data. The sample strategy data is the historical data related to the vehicle's intervention actions with the target user in the automatic driving mode.
[0073] In this embodiment, multiple sample strategy data related to intervention actions with the target user in the automatic driving mode can be obtained, and the sample data include sample strategy parameters and actual strategy parameters; the basic strategy model is trained for multiple rounds of iterations through the sample strategy parameters in the multiple sample strategy data; after each round of training, the prediction strategy parameters corresponding to the current round of training are obtained; the prediction strategy loss corresponding to the current round of training is obtained through the prediction strategy parameters corresponding to the current round of training and the actual strategy parameters in the sample strategy data corresponding to the current round of training; the basic strategy model is optimized through the prediction strategy loss corresponding to the current round of training; when the basic strategy model meets the preset conditions, the training is stopped to obtain the target strategy model. Among them, the sample strategy parameters may include sample driving parameters and sample environment parameters. The basic strategy model can also be MILE.
[0074] The trained target strategy model can take driving parameters and environmental parameters as input and output control strategies, thereby simulating the target user's vehicle control strategies under different circumstances. The target control strategy can be obtained by processing the vehicle driving parameters and environmental parameters through the target strategy model.
[0075] Based on the manual intervention of the target user in the autonomous driving mode, for example, suddenly controlling the steering wheel or changing the vehicle speed, the driving parameters and environmental parameters of these events can be recorded, such as contextual information, including but not limited to speed, position coordinates of surrounding vehicles, road conditions, and vehicle status before and after the intervention.
[0076] Using machine learning algorithms, the vehicle can analyze the patterns of these interventions and identify possible driving preferences or dissatisfaction with the performance of the automated system. In this way, the system can adjust the automated driving strategy to better simulate the user's driving habits and reduce the need for future interventions.
[0077] Let the intervention action be An, the vehicle states before and after the intervention be S_pre and S_post, and the relevant context information, such as speed V, surrounding vehicle position coordinates P, and road conditions C, be Context. The vehicle adjustment parameters can be expressed as a function f(An, S_pre, S_post, V, P, Context).
[0078] The target control strategy P_auto^n can be transformed by a function of the following form:
[0079] Among them, α is the learning rate, It represents the gradient of the relevant parameters to the intervention behavior function, and P_auto is the current control strategy.
[0080] In a possible implementation, when the target action is multiple actions, the target control parameters corresponding to each target action can be determined separately, and the target control parameters corresponding to all target actions can be determined as total control parameters, and then the vehicle is controlled by the total control parameters to achieve multiple target actions in the target driving scenario.
[0081] Figure 3 is a block diagram of a vehicle control device according to an exemplary embodiment. Figure 3 The vehicle control device 300 includes a determination module 301 , a first acquisition module 302 and a control module 303 .
[0082] The determination module 301 is configured to determine a target action corresponding to a target driving scene; The first obtaining module 302 is configured to process the target data through the target action model corresponding to the target action to obtain the target control parameter, wherein the target data is data related to executing the target action, and the target action model is obtained by training a basic model based on a plurality of sample data, and the sample data is historical data related to the target action during the historical driving process of the target user; The control module 303 is configured to control the vehicle according to the target control parameter.
[0083] Optionally, the vehicle control device 300 further includes: An acquisition module is configured to acquire a plurality of sample data related to the target action during the historical driving process of the target user, wherein the sample data includes sample association parameters and actual control parameters; A training module, configured to perform multiple rounds of iterative training on the basic model using sample-associated parameters in the multiple sample data; The second acquisition module is configured to obtain the prediction control parameters corresponding to the current round of training after each round of training; A third obtaining module is configured to obtain the prediction loss corresponding to the current round of training through the prediction control parameters corresponding to the current round of training and the actual control parameters in the sample data corresponding to the current round of training; An optimization module, configured to optimize the basic model by using the prediction loss corresponding to the current round of training; The fourth acquisition module is configured to stop training and obtain the target action model when the basic model meets the preset conditions.
[0084] Optionally, the target action includes a turn signal operation, the target action model includes a turn signal model, and the target data includes a target road condition and a target driving scene; The first obtaining module 302 includes: The first acquisition submodule is configured to process the target road condition and the target driving scenario through the turn signal model to obtain the turn signal control parameters corresponding to the turn signal operation, wherein the turn signal model is obtained by training a basic turn signal model based on multiple sample turning data, and the sample turning data are sample road conditions, sample driving scenarios and actual turn signal parameters related to the turn signal operation during the historical driving process of the target user.
[0085] Optionally, the target action includes pedal operation, the target action model includes a pedal model, and the target data includes a target speed and a target traffic condition; The first obtaining module 302 includes: The second acquisition submodule is configured to process the target speed and the target traffic condition through the pedal model to obtain the pedal control parameters corresponding to the pedal operation, wherein the pedal model is obtained by training a basic pedal model based on multiple sample pedal data, and the sample pedal data are sample speeds, sample traffic conditions and actual pedal parameters related to the pedal operation during the historical driving process of the target user.
[0086] Optionally, the target action includes a speed adjustment operation, the target action model includes a speed adjustment model, and the target data includes a target road section and a target traffic condition; The first obtaining module 302 includes: The third acquisition submodule is configured to process the target road section and the target traffic condition through the speed adjustment model to obtain the speed control parameter corresponding to the speed adjustment operation, wherein the speed adjustment model is obtained by training a basic speed model based on a plurality of sample speed data, and the sample speed data are sample road sections, sample traffic conditions and actual speed parameters related to the speed adjustment operation during the historical driving process of the target user.
[0087] Optionally, the target action includes a following distance adjustment operation, the target action model includes a following distance adjustment model, and the target data includes a target speed and a target traffic condition; The first obtaining module 302 includes: The fourth acquisition submodule is configured to process the target speed and the target traffic condition through the following distance adjustment model to obtain the following distance control parameter corresponding to the following distance adjustment operation, wherein the following distance adjustment model is obtained by training a basic distance model based on multiple sample distance data, and the sample distance data are sample speeds, sample traffic conditions and actual distance parameters related to the following distance adjustment operation during the target user's historical driving process.
[0088] Optionally, the target action includes an obstacle avoidance operation, the target action model includes an obstacle avoidance model, and the target data includes a target obstacle and a target vehicle state; The first obtaining module 302 includes: The fifth acquisition submodule is configured to process the target obstacle and the target vehicle state through the obstacle avoidance model to obtain an obstacle avoidance trajectory corresponding to the obstacle avoidance operation. The obstacle avoidance model is obtained by training a basic obstacle avoidance model based on multiple sample obstacle avoidance data. The sample obstacle avoidance data are sample obstacles, sample vehicle states and actual obstacle avoidance trajectories related to the obstacle avoidance operation during the target user's historical driving process.
[0089] Optionally, the determining module 301 includes: A first determination submodule is configured to determine a target control strategy according to vehicle driving parameters and environmental parameters; A second determination submodule is configured to determine the target driving scenario according to the target control strategy, wherein the target driving scenario includes one of overtaking and lane changing, lane keeping and obstacle avoidance; The third determination submodule is configured to determine the target action corresponding to the target driving scene based on the association relationship between the driving scene and the execution action, and one target driving scene corresponds to at least one target action.
[0090] Optionally, the first determining submodule includes: The obtaining unit is configured to process the vehicle driving parameters and the environmental parameters through a target strategy model to obtain the target control strategy, wherein the target strategy model is obtained by training a basic strategy model based on multiple sample strategy data, and the sample strategy data is historical data related to the vehicle's intervention action with the target user in the automatic driving mode.
[0091] Regarding the vehicle control device 300 in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.
[0092] The present disclosure also provides a computer-readable storage medium having computer program instructions stored thereon, and the program instructions, when executed by a processor, implement the steps of the vehicle control method provided by the present disclosure.
[0093] Figure 4 4 is a block diagram of a vehicle according to an exemplary embodiment. For example, vehicle 400 may be a hybrid vehicle, a non-hybrid vehicle, an electric vehicle, a fuel cell vehicle, or other types of vehicles. Vehicle 400 may be an autonomous vehicle, a semi-autonomous vehicle, or a non-autonomous vehicle.
[0094] Reference Figure 4 , the vehicle 400 may include various subsystems, for example, an infotainment system 410, a perception system 420, a decision control system 430, a drive system 440, and a computing platform 450. The vehicle 400 may also include more or fewer subsystems, and each subsystem may include multiple components. In addition, each subsystem and each component of the vehicle 400 may be interconnected by wire or wireless means.
[0095] In some embodiments, the infotainment system 410 may include a communication system, an entertainment system, and a navigation system, among others.
[0096] The perception system 420 may include several sensors for sensing information about the environment around the vehicle 400. For example, the perception system 420 may include a global positioning system (the global positioning system may be a GPS system, or a Beidou system or other positioning systems), an inertial measurement unit (IMU), a laser radar, a millimeter wave radar, an ultrasonic radar, and a camera.
[0097] The decision control system 430 may include a computing system, a vehicle controller, a steering system, a throttle, and a braking system.
[0098] The drive system 440 may include components that provide power movement for the vehicle 400. In one embodiment, the drive system 440 may include an engine, an energy source, a transmission system, and wheels. The engine may be one or a combination of an internal combustion engine, an electric motor, and an air compression engine. The engine is capable of converting energy provided by the energy source into mechanical energy.
[0099] Some or all functions of the vehicle 400 are controlled by a computing platform 450. The computing platform 450 may include at least one processor 451 and a memory 452, and the processor 451 may execute instructions 453 stored in the memory 452.
[0100] The processor 451 may be any conventional processor, such as a commercially available CPU. The processor may also include a graphics processor (Graphic Process Unit, GPU), a field programmable gate array (Field Programmable Gate Array, FPGA), a system on chip (System on Chip, SOC), an application specific integrated circuit (Application Specific Integrated Circuit, ASIC) or a combination thereof.
[0101] The memory 452 may be implemented by any type of volatile or nonvolatile memory device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk, or optical disk.
[0102] In addition to instructions 453 , memory 452 may also store data, such as road maps, route information, and data such as the location, direction, and speed of the vehicle. The data stored in memory 452 may be used by computing platform 450 .
[0103] In the disclosed embodiment, the processor 451 may execute the instruction 453 to complete all or part of the steps of the above-mentioned vehicle control method.
[0104] In another exemplary embodiment, a computer program product is also provided. The computer program product includes a computer program executable by a programmable device. The computer program has a code portion for executing the above-mentioned vehicle control method when executed by the programmable device.
[0105] Those skilled in the art may also understand that the various illustrative logical blocks and steps listed in the embodiments of the present application may be implemented by electronic hardware, computer software, or a combination of the two. Whether such functions are implemented by hardware or software depends on the specific application and the design requirements of the entire system. Those skilled in the art may use various methods to implement the functions described for each specific application, but such implementation should not be understood as exceeding the scope of protection of the embodiments of the present application.
[0106] It should be understood that the features of some embodiments of the various present disclosures described herein may be combined with each other, unless specifically stated otherwise.
[0107] Although terms such as "first", "second" and "third" can be used herein to describe various components, parts, regions, layers or sections, these components, parts, regions, layers or sections are not limited to these terms. On the contrary, these terms are only used to distinguish one component, component, region, layer or section from another component, component, region, layer or section. Therefore, without departing from the teachings of each example, the first component, component, region, layer or section mentioned in the examples described herein may also be referred to as the second component, component, region, layer or section. In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features. Thus, the features defined as "first" and "second" may expressly or implicitly include at least one of the features. In the description herein, the meaning of "multiple" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.
[0108] In addition, the word "exemplary" is used herein to indicate serving as an example, instance, or diagram. Any aspect or design described as "exemplary" in this article is not necessarily understood to be advantageous compared to other aspects or designs. On the contrary, the use of the word exemplary is intended to present concepts in a specific way. As used herein, the term "or" is intended to represent an inclusive "or" rather than an exclusive "or". That is, unless otherwise specified or clear from the context, "X applies A or B" is intended to represent any one of the natural inclusive arrangements. That is, if X applies A; X applies B; or X applies both A and B, "X applies A or B" is satisfied under any of the aforementioned examples. In addition, unless otherwise specified or clearly pointed to a singular form from the context, the articles "one" and "an" as used in this application and the appended claims are generally understood to mean "one or more".
[0109] Likewise, although the present disclosure has been shown and described with respect to one or more implementations, equivalent variations and modifications will occur to those skilled in the art after reading and understanding the specification and drawings. The present disclosure includes all such modifications and variations and is limited only by the scope of the claims. In particular, with respect to the various functions performed by the components (e.g., elements, resources, etc.) described above, unless otherwise indicated, the terms used to describe such components are intended to correspond to any component (functionally equivalent) that performs the specific functions of the described components, even if the structure is not equivalent to the disclosed structure. In addition, although specific features of the present disclosure may have been disclosed with respect to only one of several implementations, such features may be combined with one or more other features of other implementations as may be desired and beneficial to any given or specific application. In addition, with respect to "including", "having", "having", "having", or variations thereof used in a specific embodiment or claim, such terms are intended to be inclusive in a manner similar to the term "comprising".
[0110] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the invention disclosed herein. This application is intended to cover any modification, use or adaptation of the present disclosure, which follows the general principles of the present disclosure and includes common knowledge or customary techniques in the art that are not disclosed in the present disclosure. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present disclosure are indicated by the appended claims.
[0111] It should be understood that the present disclosure is not limited to the exact structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.
Claims
1. A vehicle control method, characterized in that: include: Determine the target action corresponding to the target driving scenario; The target data is processed by a target action model corresponding to the target action to obtain a target control parameter, wherein the target data is data related to executing the target action, and the target action model is obtained by training a basic model based on a plurality of sample data, and the sample data is historical data related to the target action during the historical driving process of the target user; The vehicle is controlled according to the target control parameter.
2. The vehicle control method according to claim 1, characterized in that: The target action model is obtained by the following steps: Acquire a plurality of sample data related to the target action during the target user's historical driving process, wherein the sample data includes sample association parameters and actual control parameters; Performing multiple rounds of iterative training on the basic model using sample association parameters in the multiple sample data; After each round of training, the prediction control parameters corresponding to this round of training are obtained; The prediction loss corresponding to this round of training is obtained through the prediction control parameters corresponding to this round of training and the actual control parameters in the sample data corresponding to this round of training; The basic model is optimized through the prediction loss corresponding to this round of training; When the basic model meets the preset conditions, the training is stopped to obtain the target action model.
3. The vehicle control method according to claim 1, characterized in that: The target action includes a turn signal operation, the target action model includes a turn signal model, and the target data includes a target road condition and a target driving scene; The target data is processed by the target action model corresponding to the target action to obtain the target control parameter, including: The target road condition and the target driving scenario are processed by the turn signal model to obtain turn signal control parameters corresponding to the turn signal operation. The turn signal model is obtained by training a basic turn signal model based on multiple sample turning data. The sample turning data are sample road conditions, sample driving scenarios and actual turn signal parameters related to the turn signal operation during the historical driving process of the target user.
4. The vehicle control method according to claim 1, characterized in that: The target action includes pedal operation, the target action model includes a pedal model, and the target data includes a target speed and a target traffic condition; The target data is processed by the target action model corresponding to the target action to obtain the target control parameter, including: The target speed and the target traffic condition are processed by the pedal model to obtain pedal control parameters corresponding to the pedal operation. The pedal model is obtained by training a basic pedal model based on multiple sample pedal data. The sample pedal data are sample speeds, sample traffic conditions and actual pedal parameters related to the pedal operation during the historical driving process of the target user.
5. The vehicle control method according to claim 1, characterized in that: The target action includes a speed adjustment operation, the target action model includes a speed adjustment model, and the target data includes a target road section and a target traffic condition; The target data is processed by the target action model corresponding to the target action to obtain the target control parameter, including: The target road section and the target traffic condition are processed by the speed adjustment model to obtain a speed control parameter corresponding to the speed adjustment operation. The speed adjustment model is obtained by training a basic speed model based on a plurality of sample speed data, and the sample speed data are sample road sections, sample traffic conditions and actual speed parameters related to the speed adjustment operation during the historical driving process of the target user.
6. The vehicle control method according to claim 1, characterized in that: The target action includes a following distance adjustment operation, the target action model includes a following distance adjustment model, and the target data includes a target speed and a target traffic condition; The target data is processed by the target action model corresponding to the target action to obtain the target control parameter, including: The target speed and the target traffic condition are processed by the following distance adjustment model to obtain a following distance control parameter corresponding to the following distance adjustment operation. The following distance adjustment model is obtained by training a basic distance model based on a plurality of sample distance data, and the sample distance data are sample speeds, sample traffic conditions and actual distance parameters related to the following distance adjustment operation during the historical driving process of the target user.
7. The vehicle control method according to claim 1, characterized in that: The target action includes an obstacle avoidance operation, the target action model includes an obstacle avoidance model, and the target data includes a target obstacle and a target vehicle state; The target data is processed by the target action model corresponding to the target action to obtain the target control parameter, including: The target obstacle and the target vehicle state are processed by the obstacle avoidance model to obtain an obstacle avoidance trajectory corresponding to the obstacle avoidance operation. The obstacle avoidance model is obtained by training a basic obstacle avoidance model based on multiple sample obstacle avoidance data. The sample obstacle avoidance data are sample obstacles, sample vehicle states and actual obstacle avoidance trajectories related to the obstacle avoidance operation during the target user's historical driving process.
8. The vehicle control method according to any one of claims 1 to 7, characterized in that: The determining of the target action corresponding to the target driving scene includes: Determine the target control strategy according to vehicle driving parameters and environmental parameters; Determining the target driving scenario according to the target control strategy; Based on the association relationship between the driving scene and the execution action, the target action corresponding to the target driving scene is determined, and one target driving scene corresponds to at least one target action.
9. The vehicle control method according to claim 8, characterized in that: Determining the target control strategy according to the vehicle driving parameters and the environmental parameters includes: The target control strategy is obtained by processing the vehicle driving parameters and the environmental parameters through a target strategy model, and the target strategy model is obtained by training a basic strategy model based on multiple sample strategy data, and the sample strategy data is historical data related to the vehicle's intervention actions with the target user in the automatic driving mode.
10. A vehicle control device, characterized in that: include: A determination module, configured to determine a target action corresponding to a target driving scenario; A first obtaining module is configured to process target data through a target action model corresponding to the target action to obtain a target control parameter, wherein the target data is data related to executing the target action, and the target action model is obtained by training a basic model based on a plurality of sample data, and the sample data is historical data related to the target action during the historical driving process of the target user; The control module is configured to control the vehicle according to the target control parameter.
11. A vehicle, characterized in that: include: processor; a memory for storing processor-executable instructions; Wherein, the processor is configured to implement the steps of the vehicle control method described in any one of claims 1 to 9 when executed.
12. A computer-readable storage medium having computer program instructions stored thereon, characterized in that: When the computer program instructions are executed by a processor, the steps of the vehicle control method described in any one of claims 1 to 9 are implemented.
13. A computer program product, characterized in that It includes a computer program, which, when executed by a processor, implements the steps of the vehicle control method described in any one of claims 1 to 9.
Citation Information
Patent Citations
Automatic driving method and device
CN111824157A
Vehicle driving assistance system, vehicle using the same, and corresponding method and medium
CN112918472A
Driving strategy acquisition method and device, computer equipment and storage medium
CN113085873A
Man-machine co-driving intention fusion control method, device and equipment and storage medium
CN116198534A
Ramp driving decision control method, vehicle and storage medium
CN117261895A
