A method, apparatus, device, and medium for determining control decisions of autonomous vehicles.

By collecting and processing environmental data of historical vehicles around autonomous vehicles, and training control decision models, the problem of poor scene adaptability of autonomous vehicles is solved, and the prediction accuracy is improved.

CN115848413BActive Publication Date: 2026-11-17GUANGZHOU WERIDE TECH LTD CO
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
CN202211673532.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-12-26
Publication Date
2026-11-17
Estimated Expiration
2042-12-26

AI Technical Summary

Technical Problem

Existing methods for determining control decisions in autonomous vehicles have poor scenario adaptability, resulting in low prediction accuracy.

Method used

The system collects current environmental data of autonomous vehicles and obtains historical environmental data of participating vehicles within a preset range, centered on the vehicle. After preprocessing, it generates training data to train the control decision model. The current environmental data is then input to obtain control decisions.

Benefits of technology

It improves the autonomous vehicle's ability to understand the current scene and enhances the predictive accuracy of control decisions.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115848413B_ABST
    Figure CN115848413B_ABST
Patent Text Reader

Abstract

This invention discloses a method, apparatus, device, and medium for determining control decisions for autonomous vehicles, addressing the technical problem of poor scenario adaptability and low prediction accuracy in existing control decision determination methods for autonomous vehicles. The invention includes: collecting current environmental data of the autonomous vehicle; acquiring historical environmental data of each historically participating vehicle within a preset range, centered on the autonomous vehicle; preprocessing the historical environmental data to obtain training data; training a control decision model using the training data; and inputting the current environmental data into the control decision model to obtain a control decision.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of autonomous driving control technology, and in particular to a method, apparatus, device, and medium for determining control decisions for autonomous vehicles. Background Technology

[0002] Autonomous vehicles, also known as driverless cars, computer-driven cars, or wheeled mobile robots, are intelligent vehicles that achieve driverless operation through computer systems. These vehicles rely on the collaborative efforts of artificial intelligence, computer vision, radar, monitoring devices, and a global positioning system to enable computers to automatically and safely operate motor vehicles without any active human intervention.

[0003] Currently, the control of autonomous driving is either an end-to-end control algorithm that takes an image signal as input and directly outputs control signals for the steering wheel, accelerator, and brake, or the entire system is a black box. Although it can achieve certain functions in some simulation scenarios, it cannot be implemented and promoted because it has problems that are difficult to explain, debug, and debug in actual engineering. In addition, its poor understanding of the scenario leads to low prediction accuracy. Summary of the Invention

[0004] This invention provides a method, apparatus, device, and medium for determining control decisions for autonomous vehicles, which addresses the technical problem of poor scenario adaptability and low prediction accuracy in existing methods for determining control decisions for autonomous vehicles.

[0005] This invention provides a method for determining control decisions for autonomous vehicles, applied to autonomous vehicles; the method includes:

[0006] Collect current environmental data of the autonomous vehicle;

[0007] Centered on the autonomous vehicle, acquire historical environmental data of each historically participating vehicle within a preset range;

[0008] The historical environmental data is preprocessed to obtain training data;

[0009] The control decision model is trained using the training data.

[0010] The current environmental data is input into the control decision model to obtain a control decision.

[0011] Optionally, after the step of inputting the current environmental data into the control decision model to obtain the control decision, the method further includes:

[0012] The driving behavior corresponding to the control decision is executed.

[0013] Optionally, the step of acquiring historical environmental data of each historically participating vehicle within a preset range, centered on the autonomous vehicle, includes:

[0014] Centered on the autonomous vehicle, acquire historical participating vehicles within a preset range and a preset time range;

[0015] Acquire historical environmental data collected from each participating vehicle.

[0016] Optionally, the step of preprocessing the historical environmental data to obtain training data includes:

[0017] Determine the scene type corresponding to the current environmental data;

[0018] Target scene data is filtered from the historical environmental data according to the scene type;

[0019] Generate the trajectory data of the historical participating vehicles based on the target scene data;

[0020] The tag information of the historically participating vehicles is generated based on the trajectory data and the target scene data;

[0021] The target scene data and the label information are used as training data.

[0022] Optionally, the target scene data includes obstacle information; the step of generating the tag information of the historically participating vehicles based on the trajectory data and the target scene data includes:

[0023] Based on the trajectory data and the obstacle information, the relative behavior of the historically participating vehicles with respect to the obstacle information is determined;

[0024] The absolute behavior of the historical participating vehicles relative to the ground is determined based on the trajectory data;

[0025] The tag information of the historically participating vehicles is generated based on the relative behavior and the absolute behavior.

[0026] The present invention also provides an autonomous vehicle control decision determination device, applied to an autonomous vehicle; the device includes:

[0027] The current environment data acquisition module is used to collect the current environment data of the autonomous vehicle;

[0028] The historical environment data acquisition module is used to acquire historical environment data of each historically participating vehicle within a preset range, with the autonomous vehicle as the center.

[0029] The training data acquisition module is used to preprocess the historical environment data to obtain training data;

[0030] A control decision model training module is used to train a control decision model using the training data.

[0031] The control decision generation module is used to input the current environmental data into the control decision model to obtain control decisions.

[0032] Optionally, the device further includes:

[0033] The execution module is used to execute the driving behavior corresponding to the control decision.

[0034] Optionally, the historical environmental data acquisition module includes:

[0035] The historical participating vehicle acquisition submodule is used to acquire historical participating vehicles within a preset range and a preset time range, with the autonomous vehicle as the center.

[0036] The historical environmental data acquisition submodule is used to acquire historical environmental data collected by each participating vehicle.

[0037] The present invention also provides an electronic device, the device comprising a processor and a memory:

[0038] The memory is used to store program code and transmit the program code to the processor;

[0039] The processor is configured to execute the autonomous vehicle control decision determination method as described above, according to instructions in the program code.

[0040] The present invention also provides a computer-readable storage medium for storing program code for performing the autonomous vehicle control decision determination method as described in any of the preceding claims.

[0041] As can be seen from the above technical solutions, the present invention has the following advantages: The present invention provides a method for determining control decisions for autonomous vehicles, and specifically discloses: collecting current environmental data of autonomous vehicles; obtaining historical environmental data of each historically participating vehicle within a preset range, with the autonomous vehicle as the center; preprocessing the historical environmental data to obtain training data; training the control decision model using the training data; and inputting the current environmental data into the control decision model to obtain control decisions.

[0042] This invention uses historical environmental data of participating vehicles within a preset range of the current autonomous vehicle's environment as a basis to extract training data. A control decision model is then trained based on this data, enabling the autonomous vehicle to generate control decisions for the current road segment and scenario based on the behavioral data of historical participating vehicles. Because this control decision model is derived from the historical environmental data of historical participating vehicles in the current road segment, it has a stronger understanding of the current scenario, thereby improving the prediction accuracy of control decisions. Attached Figure Description

[0043] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0044] Figure 1 A flowchart illustrating the steps of a method for determining control decisions for autonomous vehicles, as provided in an embodiment of the present invention;

[0045] Figure 2 A flowchart illustrating the steps of a method for determining control decisions for autonomous vehicles, as provided in another embodiment of the present invention;

[0046] Figure 3 This is a structural block diagram of an autonomous vehicle control decision determination device provided in an embodiment of the present invention. Detailed Implementation

[0047] This invention provides a method, apparatus, device, and medium for determining control decisions for autonomous vehicles, which addresses the technical problem of poor scenario adaptability and low prediction accuracy in existing methods for determining control decisions for autonomous vehicles.

[0048] To make the objectives, features, and advantages of this invention more apparent and understandable, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described below are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0049] Please see Figure 1 , Figure 1 A flowchart illustrating the steps of a method for determining control decisions for autonomous vehicles, as provided in an embodiment of the present invention.

[0050] This invention provides a method for determining control decisions for autonomous vehicles, applicable to autonomous vehicles; the method may specifically include the following steps:

[0051] Step 101: Collect current environmental data of the autonomous vehicle;

[0052] Autonomous vehicles, also known as driverless cars, computer-driven cars, or wheeled mobile robots, are intelligent vehicles that achieve driverless operation through computer systems. These vehicles rely on the collaborative efforts of artificial intelligence, computer vision, radar, monitoring devices, and a global positioning system to enable computers to automatically and safely operate motor vehicles without any active human intervention.

[0053] The current environmental data can be data collected automatically (or manually controlled by a safety operator) by the data acquisition device around the autonomous vehicle (the collection range can be set according to actual needs, such as within a radius of 50 meters; this embodiment of the invention does not specifically limit the collection range). This data includes the shape, position, speed, and acceleration of obstacles, pre-annotated semantic map data in the scene, and the status of surrounding traffic control equipment. Due to the large amount of data, the collected information can be stored on storage media such as hard drives on the autonomous vehicle.

[0054] Step 102: Taking the autonomous vehicle as the center, acquire the historical environmental data of each historically participating vehicle within a preset range;

[0055] In this embodiment of the invention, the preset range can be the complete road segment where the current autonomous vehicle is located, or it can be a pre-set radius range. This embodiment of the invention does not specifically limit this range.

[0056] Before making control decisions based on the current environmental data it collects, autonomous vehicles can acquire historical environmental data of all participating vehicles within a preset range centered on the autonomous vehicle to generate a control decision model suitable for the current scenario. The control decision model then generates the corresponding control decisions.

[0057] Among them, historical participating vehicles refer to vehicles that appeared within the aforementioned preset range during the preset time period and collected and uploaded environmental data, including but not limited to autonomous vehicles, semi-autonomous vehicles and non-autonomous vehicles of various models.

[0058] Historical environmental data of historical participating vehicles refers to the shape, position, speed, and acceleration of obstacles identified by historical participating vehicles within the aforementioned preset range, semantic map data pre-annotated in the scene, the status of surrounding traffic control equipment, as well as the vehicle model, position, speed, and acceleration of the historical participating vehicles themselves over time.

[0059] The historical environmental data of the participating vehicles is extensive and complex. Therefore, the data collected from each participating vehicle can be uploaded to the cloud for centralized storage, which will facilitate subsequent model training.

[0060] It should be noted that due to the large amount of historical environmental data, in order to ensure data validity and avoid wasting storage space, it is necessary to filter and remove data that does not meet the requirements. Data to be removed may include data collected when the system is unstable, such as data collected during hardware failures, data collected when modules are incomplete, data from environments that have not changed for a long time, and unreliable data due to weather or signal issues.

[0061] Step 103: Preprocess the historical environmental data to obtain training data;

[0062] After obtaining historical environmental data of the vehicles involved, it can be preprocessed to obtain training data for the control decision model.

[0063] By utilizing historical environmental data, a holistic 3D reconstruction with a time dimension can be performed on the entire scene involving the vehicles. This reveals the observable position and motion state of each participant at a given moment, including their position, orientation, speed, acceleration, and angular velocity. Details such as vehicle turn signals, taillights, reversing lights, road condition camera status, pedestrian speed, and facial orientation can also be obtained. However, in autonomous driving control decisions, only the vehicle's behavior (different vehicle behaviors correspond to different control decisions) needs to be categorized. Therefore, this invention requires further filtering of the historical environmental data to obtain training data for the control decision model.

[0064] In one example, different vehicle behaviors can be determined by the historical relative behavior of participating vehicles with respect to other vehicles and their absolute behavior with respect to the ground. Specific control decisions may include:

[0065] Change lanes to the left: Complete a left lane change relative to the map;

[0066] Turn left: Move out of the current lane relative to the vehicle in front and overtake it in the longitudinal direction;

[0067] Abandoning the lane change: The vehicle made a relatively large crossing of the lane line, but eventually returned to its current lane;

[0068] Keep straight: Do not cross any obstacles in your lane within a certain time period.

[0069] Step 104: Train the control decision model using training data;

[0070] After obtaining the training data, the control decision model can be trained using the training data. The control decision model can be obtained through conventional deep learning training. This invention does not impose specific restrictions on the training process.

[0071] Step 105: Input the current environmental data into the control decision model to obtain the control decision.

[0072] After training the control decision model, current environmental data can be input into the model to obtain control decisions for the current scenario, such as going straight, detouring to the left, or changing lanes to the left.

[0073] This invention uses historical environmental data of participating vehicles within a preset range of the current autonomous vehicle's environment as a basis to extract training data. A control decision model is then trained based on this data, enabling the autonomous vehicle to generate control decisions for the current road segment and scenario based on the behavioral data of historical participating vehicles. Because this control decision model is derived from the historical environmental data of historical participating vehicles in the current road segment, it has a stronger understanding of the current scenario, thereby improving the prediction accuracy of control decisions.

[0074] Please see Figure 2 , Figure 2 A flowchart illustrating the steps of a method for determining control decisions for an autonomous vehicle, as provided in another embodiment of the present invention. Specifically, it may include the following steps:

[0075] Step 201: Collect current environmental data of the autonomous vehicle;

[0076] Step 201 is the same as step 101. For details, please refer to the description of step 101. It will not be repeated here.

[0077] Step 202: Taking the autonomous vehicle as the center, acquire the historical environmental data of each historically participating vehicle within a preset range;

[0078] In this embodiment of the invention, before making a control decision based on the current environmental data collected by the autonomous vehicle, the autonomous vehicle can acquire historical environmental data of each historically participating vehicle within a preset range centered on the autonomous vehicle to generate a control decision model suitable for the current scenario, and then the control decision model generates the corresponding control decision.

[0079] In one example, the step of acquiring historical environmental data of each historically participating vehicle within a preset range, centered on the autonomous vehicle, may include the following sub-steps:

[0080] S21, with autonomous vehicles as the center, acquires historical participating vehicles within a preset range and time range;

[0081] S22, acquire historical environmental data collected by each participating vehicle.

[0082] In this embodiment of the invention, historical participating vehicles within a preset range and time range centered on the autonomous vehicle can be obtained from the cloud. For example, historical participating vehicles that appeared within a 100-meter radius centered on the autonomous vehicle within one month prior to the current time of the autonomous vehicle.

[0083] Next, retrieve the relevant historical environmental data from the cloud.

[0084] Step 203: Preprocess the historical environmental data to obtain training data;

[0085] After obtaining historical environmental data of the vehicles involved, it can be preprocessed to obtain training data for the control decision model.

[0086] In one example, the step of preprocessing historical environmental data to obtain training data may include the following sub-steps:

[0087] S31, Determine the scenario type corresponding to the current environmental data;

[0088] S32, filter target scene data from historical environmental data according to scene type;

[0089] S33, Generate trajectory data of historically involved vehicles based on target scene data;

[0090] S34, Generate tag information of historically involved vehicles based on trajectory data and target scene data;

[0091] In this embodiment of the invention, the historical environmental data of the participating vehicles contains a lot of information. However, not all historical environmental data is applicable to the current scenario faced by the autonomous vehicle. Therefore, in this embodiment of the invention, it is first necessary to determine the scenario type corresponding to the current environmental data in order to filter target scenario data that meets the requirements of the current scenario.

[0092] The target scene data is recorded as data collected by historically participating vehicles within a continuous timeframe in the current scene. Therefore, trajectory data of historically participating vehicles can be generated according to the timeline. By combining the trajectory data with the driving situation of obstacles in front of the corresponding vehicles, the driving behavior taken by the vehicles can be determined, and the vehicle's tag information can be obtained based on this driving behavior.

[0093] Specifically, the step of generating tag information for historically participating vehicles based on trajectory data and target scene data may include the following sub-steps:

[0094] S341, Based on trajectory data and obstacle information, determine the relative behavior of historically participating vehicles with respect to obstacle information;

[0095] S342, determine the absolute behavior of historical participating vehicles relative to the ground based on trajectory data;

[0096] S343, Generate tag information for historically involved vehicles based on relative and absolute behaviors.

[0097] In practical implementation, the relative behavior between historical participating vehicles and obstacles, as well as the absolute behavior between historical participating vehicles and the ground, can be obtained from the target scene data to determine the driving behavior of historical participating vehicles, and thus generate label information for the driving behavior of historical participating vehicles based on the driving behavior.

[0098] For example, if a vehicle in the past has completed a lane change to the left relative to the map (ground), its label information can be "lane change to the left".

[0099] If a historically participating vehicle moves out of its current lane relative to the vehicle in front and completes overtaking the vehicle in the longitudinal direction, its label information can be "detour to the left".

[0100] If a vehicle that participated in the previous event made a relatively large crossing of the lane line (such as the angle between the vehicle body and the lane line being greater than a preset angle, such as 5 degrees or 10 degrees, which can be set according to the actual situation), but eventually returned to the current lane, the tag information can be "abandon lane change".

[0101] If a vehicle has not crossed any obstacle in its lane within a certain period of time, its tag information can be "Keep Straight".

[0102] It should be noted that the tag information corresponding to the driving behavior of historically involved vehicles can also be manually determined by management personnel.

[0103] S35 uses target scene data and label information as training data.

[0104] After obtaining the target scene data and the corresponding label information, the target scene data and label data can be used as training data.

[0105] Step 204: Train the control decision model using training data;

[0106] After obtaining the training data, the control decision model can be trained using the training data; the control decision model can be obtained through conventional deep learning training, and this invention does not impose specific restrictions on the training process.

[0107] It should be noted that different control decision models can be trained according to different scenarios. This invention does not impose specific restrictions on the type of control decision model or the training process. Those skilled in the art can flexibly set it based on conventional deep learning and actual scenario requirements.

[0108] It should be noted that, considering the large amount of historical environmental data from participating vehicles, the cloud may not necessarily send the relevant historical environmental data to the autonomous vehicle after filtering it out. Instead, the control decision model can be trained in the cloud and then sent to the autonomous vehicle.

[0109] Step 205: Input the current environmental data into the control decision model to obtain the control decision;

[0110] After the control decision model is generated, the input data required by the control decision model can be extracted from the current environmental data and input into the control decision model to obtain the corresponding control decision.

[0111] Step 206: Execute the driving behavior corresponding to the control decision.

[0112] After obtaining the control decision, the corresponding driving behavior can be executed, such as changing lanes to the left, detouring to the left, or continuing straight.

[0113] This invention uses historical environmental data of participating vehicles within a preset range of the current autonomous vehicle's environment as a basis to extract training data. A control decision model is then trained based on this data, enabling the autonomous vehicle to generate control decisions for the current road segment and scenario based on the behavioral data of historical participating vehicles. Because this control decision model is derived from the historical environmental data of historical participating vehicles in the current road segment, it has a stronger understanding of the current scenario, thereby improving the prediction accuracy of control decisions.

[0114] Please see Figure 3 , Figure 3 This is a structural block diagram of an autonomous vehicle control decision determination device provided in an embodiment of the present invention.

[0115] This invention provides an autonomous vehicle control decision determination device, applied to an autonomous vehicle; the device includes:

[0116] The current environment data acquisition module 301 is used to collect current environment data of autonomous vehicles;

[0117] The historical environment data acquisition module 302 is used to acquire historical environment data of each historically participating vehicle within a preset range, with the autonomous vehicle as the center.

[0118] The training data acquisition module 303 is used to preprocess historical environmental data to obtain training data;

[0119] The control decision model training module 304 is used to train the control decision model using training data.

[0120] The control decision generation module 305 is used to input the current environmental data into the control decision model to obtain control decisions.

[0121] In this embodiment of the invention, the device further includes:

[0122] The execution module is used to execute the driving behaviors corresponding to the control decisions.

[0123] In this embodiment of the invention, the historical environmental data acquisition module 302 includes:

[0124] The historical participating vehicle acquisition submodule is used to acquire historical participating vehicles within a preset range and time range, centered on autonomous vehicles.

[0125] The historical environmental data acquisition submodule is used to acquire historical environmental data collected by each participating vehicle.

[0126] In this embodiment of the invention, the training data acquisition module 303 includes:

[0127] The scenario type determination submodule is used to determine the scenario type corresponding to the current environment data;

[0128] The target scene data filtering submodule is used to filter target scene data from historical environmental data based on scene type;

[0129] The trajectory data generation submodule is used to generate trajectory data of historically participating vehicles based on the target scene data;

[0130] The tag information generation submodule is used to generate tag information for historically participating vehicles based on trajectory data and target scene data;

[0131] The training data generation submodule is used to take target scene data and label information as training data.

[0132] In this embodiment of the invention, the target scene data includes obstacle information; the tag information generation submodule includes:

[0133] The relative behavior determination unit is used to determine the relative behavior of historically participating vehicles relative to obstacle information based on trajectory data and obstacle information;

[0134] The absolute behavior determination unit is used to determine the absolute behavior of historical participating vehicles relative to the ground based on trajectory data.

[0135] The tag information generation unit is used to generate tag information for historically participating vehicles based on relative and absolute behaviors.

[0136] The present invention also provides an electronic device, the device including a processor and a memory:

[0137] The memory is used to store program code and transfer the program code to the processor;

[0138] The processor is used to execute the autonomous vehicle control decision determination method of this invention according to the instructions in the program code.

[0139] This invention also provides a computer-readable storage medium for storing program code for executing the autonomous vehicle control decision determination method of this invention.

[0140] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0141] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0142] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, apparatus, or computer program products. Therefore, embodiments of the present invention can take the form of entirely hardware embodiments, entirely software embodiments, or embodiments combining software and hardware aspects. Furthermore, embodiments of the present invention can take the form of computer program products implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0143] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0144] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0145] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal equipment, causing a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0146] Although preferred embodiments of the present invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the embodiments of the present invention.

[0147] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element.

[0148] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An automatic driving vehicle control decision determination method, characterized by, Applied to autonomous vehicles; the method includes: Collect current environmental data of the autonomous vehicle; Centered on the autonomous vehicle, acquire historical environmental data of each historically participating vehicle within a preset range; The historical environmental data is preprocessed to obtain training data; The control decision model is trained using the training data. The current environmental data is input into the control decision model to obtain the control decision; The step of preprocessing the historical environmental data to obtain training data includes: Determine the scene type corresponding to the current environmental data; Target scenario data is filtered from the historical environmental data according to the scenario type; the target scenario data records the data collected by the participating vehicles in the current scenario over a continuous period of time. Generate the trajectory data of the historical participating vehicles based on the target scene data; The tag information of the historically participating vehicles is generated based on the trajectory data and the target scene data; The target scene data and the label information are used as training data; The target scene data includes obstacle information; the step of generating the tag information of the historically participating vehicles based on the trajectory data and the target scene data includes: Based on the trajectory data and the obstacle information, the relative behavior of the historically participating vehicles with respect to the obstacle information is determined; The absolute behavior of the historical participating vehicles relative to the ground is determined based on the trajectory data; The tag information of the historically participating vehicles is generated based on the relative behavior and the absolute behavior.

2. The method of claim 1, wherein, After the step of inputting the current environmental data into the control decision model to obtain the control decision, the method further includes: The driving behavior corresponding to the control decision is executed.

3. The method according to claim 1, characterized in that, The step of acquiring historical environmental data of each historically participating vehicle within a preset range, centered on the autonomous vehicle, includes: Centered on the autonomous vehicle, acquire historical participating vehicles within a preset range and a preset time range; Acquire historical environmental data collected from each participating vehicle.

4. A device for determining control decisions for an automated vehicle, characterized in that, Applied to autonomous vehicles; the device includes: The current environment data acquisition module is used to collect the current environment data of the autonomous vehicle; The historical environment data acquisition module is used to acquire historical environment data of each historically participating vehicle within a preset range, with the autonomous vehicle as the center. The training data acquisition module is used to preprocess the historical environment data to obtain training data; A control decision model training module is used to train a control decision model using the training data. The control decision generation module is used to input the current environmental data into the control decision model to obtain control decisions; The training data acquisition module includes: The scenario type determination submodule is used to determine the scenario type corresponding to the current environment data; The target scenario data filtering submodule is used to filter target scenario data from historical environmental data based on scenario type; the target scenario data records the data collected by participating vehicles in the current scenario over a continuous period of time. The trajectory data generation submodule is used to generate trajectory data of historically participating vehicles based on the target scene data; The tag information generation submodule is used to generate tag information for historically participating vehicles based on trajectory data and target scene data; The training data generation submodule is used to take target scene data and label information as training data; The target scene data includes obstacle information; the label information generation submodule includes: The relative behavior determination unit is used to determine the relative behavior of historically participating vehicles relative to obstacle information based on trajectory data and obstacle information; The absolute behavior determination unit is used to determine the absolute behavior of historical participating vehicles relative to the ground based on trajectory data. The tag information generation unit is used to generate tag information for historically participating vehicles based on relative and absolute behaviors.

5. The apparatus according to claim 4, characterized in that, The device further includes: The execution module is used to execute the driving behavior corresponding to the control decision.

6. The apparatus according to claim 4, characterized in that, The historical environment data acquisition module includes: The historical participating vehicle acquisition submodule is used to acquire historical participating vehicles within a preset range and a preset time range, with the autonomous vehicle as the center. The historical environmental data acquisition submodule is used to acquire historical environmental data collected by each participating vehicle.

7. An electronic device, characterized in that, The device includes a processor and a memory: The memory is used to store program code and transmit the program code to the processor; The processor is used to execute the autonomous vehicle control decision determination method according to any one of claims 1-3 according to the instructions in the program code.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium is used to store program code for executing the autonomous vehicle control decision determination method according to any one of claims 1-3.

Citation Information

Patent Citations

  • Vehicle lane changing prediction method and device and computer storage medium

    CN112918478A

  • Prediction type intelligent automobile decision control method and device, vehicle and storage medium

    CN114084155A

  • Environmental vehicle behavior prediction model training method and system

    CN114742236A