Vehicle control method and apparatus, and vehicle-mounted device, vehicle and storage medium

Abstract

Provided are a vehicle control method and apparatus, and a vehicle-mounted device, a vehicle and a storage medium. The vehicle control method comprises: on the basis of environmental perception data, position data of a vehicle, and map data, determining behavior trajectory information of a target object in an environment where the vehicle is located; sending the behavior trajectory information of the target object and ego-vehicle trajectory information of the vehicle to a server, so that the server processes the behavior trajectory information of the target object and the ego-vehicle trajectory information on the basis of a target model to generate planning and decision-making information, wherein the target model is a generative pretrained transformer and is obtained by means of performing training on the basis of a plurality of types of training sample data; receiving the planning and decision-making information fed back by the server, and on the basis of the behavior trajectory information of the target object and the planning and decision-making information, generating planning trajectory information and driving strategy information for the vehicle; and on the basis of the planning trajectory information and the driving strategy information, controlling the vehicle to travel. The safety of autonomous driving and the driving and riding experience of users can be improved.

Description

Vehicle control method, device, vehicle-mounted equipment, vehicle, and storage medium

[0001] This disclosure claims priority to a Chinese patent application filed with the Patent Office of China on February 26, 2024, with application number 2024102107242 and application name “Vehicle Control Method, Device, On-board Equipment, Vehicle and Storage Medium,” the entire contents of which are incorporated by reference into this disclosure. Technical Field

[0002] The present disclosure relates to the field of autonomous driving technology, and more specifically, to a vehicle control method, apparatus, on-board equipment, vehicle, and storage medium. Background Art

[0003] Vehicle intelligence is one of the main development directions of current vehicle technology, and autonomous driving is a key technology in this process. In autonomous driving, the vehicle plans its own driving trajectory and drives according to it.

[0004] However, in practical applications, if a vehicle's planned trajectory is inappropriate, it not only affects the autonomous driving experience but can also seriously impact the safety of other vehicles and the vehicle itself. Therefore, improving the effectiveness and comfort of trajectory planning is a long-standing goal in the autonomous driving field.

[0005] Summary of the Invention

[0006] The embodiments of the present disclosure provide at least one vehicle control method, a vehicle control device, an on-board device, a vehicle, and a storage medium, which can enhance the experience of autonomous driving.

[0007] In a first aspect, an embodiment of the present disclosure provides a vehicle control method, including:

[0008] Acquiring environmental perception data of an environment in which a vehicle is located, position data of the vehicle, and map data of the environment in which the vehicle is located, and determining behavior trajectory information of a target object in the environment in which the vehicle is located based on the environmental perception data, the position data of the vehicle, and the map data;

[0009] The target object's behavior trajectory information and the vehicle's own trajectory information are sent to a server, so that the server processes the target object's behavior trajectory information and the vehicle's own trajectory information based on a trained target model to generate planning decision information; the target model is a pre-trained transformer that is obtained through self-supervised training based on multiple training sample data;

[0010] receiving the planning decision information fed back by the server, and generating planning trajectory information and driving strategy information for the vehicle based on the behavior trajectory information of the target object and the planning decision information;

[0011] The vehicle is controlled based on the planned trajectory information and the driving strategy information.

[0012] In one possible implementation, the planning decision information includes predicted trajectory information and driving instruction information for the vehicle, the trajectory length indicated by the predicted trajectory information is greater than a preset length, and the driving instruction information is used to instruct a driving behavior of the vehicle;

[0013] The generating of planning trajectory information and driving strategy information for the vehicle based on the behavior trajectory information of the target object and the planning decision information includes:

[0014] According to the behavior trajectory information of the target object, the predicted trajectory information and the driving instruction information, planned trajectory information and driving strategy information for the vehicle are generated.

[0015] In one possible implementation, the planning decision information further includes driving style information, where the driving style information is used to indicate a degree of smoothness of vehicle driving; and controlling the vehicle driving based on the planned trajectory information and the driving strategy information includes:

[0016] The vehicle is controlled to travel based on the planned trajectory information, the driving strategy information, and the driving style information.

[0017] In a possible implementation, the behavior trajectory information of the target object includes tracking trajectory information of the target object, predicted trajectory information of the target object, and predicted behavior information of the target object.

[0018] In one possible implementation, determining the behavior trajectory information of a target object in the environment in which the vehicle is located based on the environmental perception data, the vehicle position data, and the map data includes:

[0019] performing feature extraction on the image data to obtain bird's-eye view features of the image data, and determining bird's-eye view features of the map data based on the bird's-eye view features of the image data;

[0020] Based on the position data of the vehicle, fusing the bird's-eye view feature of the image data with the bird's-eye view feature of the map data to obtain a target fusion feature;

[0021] Based on the target fusion features, behavior trajectory information of the target object is determined.

[0022] In a possible implementation, the fusing of the bird's-eye view feature of the image data with the bird's-eye view feature of the map data based on the vehicle's position data to obtain a target fusion feature includes:

[0023] fusing the bird's-eye view feature of the image data with the bird's-eye view feature of the map data at a corresponding moment to obtain a single-frame fused bird's-eye view feature;

[0024] Based on the position data of the vehicle, the relative posture data of the vehicle at adjacent moments are determined, and based on the relative posture data, the multi-frame fused bird's-eye view features at the current moment and before the current moment are aligned and fused to obtain the target fusion feature.

[0025] In a possible implementation, after obtaining the target fusion feature, the method further includes:

[0026] Determine a plurality of perception information based on the target fusion feature, wherein the plurality of perception information includes at least two of the following:

[0027] Three-dimensional object perception information, three-dimensional road structure perception information, occupied space grid perception information, traffic light perception information, three-dimensional map perception information and fused positioning perception information.

[0028] In a possible implementation, generating planning trajectory information and driving strategy information for the vehicle based on the behavior trajectory information of the target object and the planning decision information includes:

[0029] Based on the behavior trajectory information of the target object, the traffic light perception information, the three-dimensional map perception information and the planning decision information, planning trajectory information and driving strategy information for the vehicle are generated.

[0030] In a second aspect, an embodiment of the present disclosure provides a vehicle control method, including:

[0031] receiving vehicle trajectory information for the target vehicle and behavior trajectory information of a target object in the environment of the target vehicle sent by the target vehicle; the behavior trajectory information of the target object is determined based on environmental perception data of the environment in which the vehicle is located, position data of the vehicle, and map data of the environment in which the vehicle is located;

[0032] The target object's behavior trajectory information and the vehicle's trajectory information are processed based on a trained target model to generate planning decision information; the target model is a pre-trained transformer that is obtained through self-supervised training based on a variety of training sample data;

[0033] The planning decision information is sent to the target vehicle, where the planning decision information is used to instruct the target vehicle to generate planning trajectory information and driving strategy information for the vehicle based on the behavior trajectory information of the target object and the planning decision information.

[0034] In one possible implementation, the target model is trained by the following steps:

[0035] Obtaining a basic network to be trained, wherein model parameters of the basic network are greater than a preset scale;

[0036] Acquire training sample data, and perform self-supervised training on the basic network based on the training sample data to obtain the trained target model; the training sample data includes at least two of the following:

[0037] Vehicle trajectory sample data, driving status sample data, weather condition sample data during vehicle driving, traffic condition sample data during vehicle driving, navigation route sample data corresponding to vehicle driving, road structure status sample data perceived by the vehicle, trajectory sample data of dynamic and static obstacles, and Internet driving video sample data.

[0038] In a third aspect, an embodiment of the present disclosure provides a vehicle control device, comprising:

[0039] an information determination module, configured to obtain environmental perception data of an environment in which a vehicle is located, position data of the vehicle, and map data of the environment in which the vehicle is located, and determine, based on the environmental perception data, position data of the vehicle, and map data, behavior trajectory information of a target object in the environment in which the vehicle is located;

[0040] a first sending module configured to send the target object's behavior trajectory information and the vehicle's own trajectory information to a server, so that the server processes the target object's behavior trajectory information and the vehicle's own trajectory information based on a trained target model to generate planning decision information; the target model is a pre-trained transformer obtained through self-supervised training based on a plurality of training sample data;

[0041] an information generation module, configured to receive the planning decision information fed back by the server, and generate planning trajectory information and driving strategy information for the vehicle based on the behavior trajectory information of the target object and the planning decision information;

[0042] A vehicle control module is used to control the vehicle driving based on the planned trajectory information and the driving strategy information.

[0043] In a fourth aspect, an embodiment of the present disclosure provides a vehicle control device, comprising:

[0044] An information receiving module is configured to receive vehicle trajectory information of a target vehicle and behavior trajectory information of a target object in the environment of the target vehicle, the behavior trajectory information of the target object being determined based on environmental perception data of the environment in which the vehicle is located, the position data of the vehicle, and map data of the environment in which the vehicle is located;

[0045] An information processing module, configured to process the target object's behavior trajectory information and the vehicle's trajectory information based on a trained target model to generate planning decision information; the target model is a pre-trained transformer obtained through self-supervised training based on a variety of training sample data;

[0046] The second sending module is used to send the planning decision information to the target vehicle, and the planning decision information is used to instruct the target vehicle to generate planning trajectory information and driving strategy information for the vehicle based on the behavior trajectory information of the target object and the planning decision information.

[0047] In the fifth aspect, an embodiment of the present disclosure provides a vehicle-mounted device, comprising: a processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor. When the vehicle-mounted device is running, the processor and the memory communicate through the bus, and when the machine-readable instructions are executed by the processor, the vehicle control method described in any possible implementation method of the first aspect above is executed.

[0048] In a sixth aspect, an embodiment of the present disclosure provides a vehicle, including a controller, wherein the controller includes:

[0049] a memory configured to store instructions; and

[0050] The processor is configured to call the instructions from the memory and to implement the vehicle control method described in any possible implementation of the first aspect when executing the instructions.

[0051] In a seventh aspect, an embodiment of the present disclosure provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, executes the vehicle control method as described in any of the possible embodiments described above.

[0052] The vehicle control method, vehicle control device, on-board equipment, vehicle, and computer-readable storage medium provided by the embodiments of the present disclosure first send the behavior trajectory information of the target object in the vehicle's environment and the vehicle's own trajectory information to the server, so that the server can generate guiding planning decision information for the vehicle based on the driving experience of the pre-trained large model (generating a pre-trained transformer). In this way, when the vehicle performs trajectory planning, it is not only based on the behavior trajectory information of the target object in the environment, but also combined with the guiding planning decision information given by the large model. This can improve the safety of the generated planning trajectory information and the applicability of the driving strategy information, thereby improving the safety of autonomous driving and the user's driving experience.

[0053] In order to make the above-mentioned objectives, features and advantages of the present disclosure more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the following briefly introduces the drawings required for use in the embodiments. The drawings herein are incorporated into and constitute a part of the specification. These drawings illustrate embodiments consistent with the present disclosure and, together with the specification, are used to illustrate the technical solutions of the present disclosure. It should be understood that the following drawings only illustrate certain embodiments of the present disclosure and should not be regarded as limiting the scope. For those of ordinary skill in the art, other relevant drawings can be obtained based on these drawings without inventive effort.

[0055] FIG1 shows a flow chart of a vehicle control method provided by an embodiment of the present disclosure;

[0056] FIG2 is a schematic diagram showing a vehicle control process provided by an embodiment of the present disclosure;

[0057] FIG3 is a schematic diagram showing another vehicle control process provided by an embodiment of the present disclosure;

[0058] FIG4 shows a flow chart of another vehicle control method provided by an embodiment of the present disclosure;

[0059] FIG5 shows a schematic diagram of a target model training process provided by an embodiment of the present disclosure;

[0060] FIG6 shows a functional module diagram of a vehicle control device provided by an embodiment of the present disclosure;

[0061] FIG7 shows a functional module diagram of another vehicle control device provided by an embodiment of the present disclosure;

[0062] FIG8 shows a functional module diagram of another vehicle control device provided by an embodiment of the present disclosure;

[0063] FIG9 shows a schematic structural diagram of a vehicle-mounted device provided by an embodiment of the present disclosure. DETAILED DESCRIPTION

[0064] In order to make the purpose, technical solutions and advantages of the embodiments of the present disclosure clearer, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are only part of the embodiments of the present disclosure, not all of the embodiments. The components of the embodiments of the present disclosure generally described and shown in the drawings herein can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present disclosure provided in the drawings is not intended to limit the scope of the disclosure for which protection is sought, but merely represents selected embodiments of the present disclosure. Based on the embodiments of the present disclosure, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of the present disclosure.

[0065] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings.

[0066] The term "and / or" herein simply describes an association relationship, indicating that three relationships can exist. For example, A and / or B can represent the existence of A alone, the simultaneous existence of A and B, and the existence of B alone. In addition, the term "at least one" herein refers to any combination of at least two of any one or more of a plurality of items. For example, "at least one of A, B, and C" can represent any one or more elements selected from the set consisting of A, B, and C.

[0067] Autonomous vehicles use onboard devices such as the ECU (Electronic Control Unit) to precisely control and analyze various vehicle components, enabling fully automated operation and unmanned driving. With the rapid development of deep learning and the in-depth study of artificial intelligence, achieving autonomous driving through end-to-end deep learning has become a major research direction in the autonomous driving field.

[0068] Generative Pretrained Transformer (GPT), also known as big model, is increasingly being applied to various practical scenarios. It has both “large-scale” and “pre-trained” properties and can be pre-trained on massive amounts of general data, which can greatly improve the generalization, versatility, and practicality of artificial intelligence. Research has found that due to the high resource requirements and inability to guarantee real-time and robustness of this big model, its application scenarios are mostly based on big models to pre-process data for autonomous driving training (such as automatic labeling and other tasks), assist in autonomous driving model training, etc., and facing the high real-time and high robustness requirements of autonomous driving and the constraints of vehicle-side resource limitations, it cannot be directly used in the vehicle-side autonomous driving system. Therefore, how to apply the superior performance of big models to autonomous driving systems to enhance the autonomous driving experience is the focus of this application research.

[0069] Based on the above research, an embodiment of the present disclosure provides a vehicle control method, which first determines the behavior trajectory information of the target object in the vehicle's environment based on the environmental perception data of the vehicle's environment, the vehicle's position data and the map data, and then sends the behavior trajectory information of the target object and the vehicle's own trajectory information to a server, so that the server processes the behavior trajectory information of the target object and the vehicle's own trajectory information based on a trained target model to generate planning decision information; then, based on the behavior trajectory information of the target object and the planning decision information fed back by the server, generates planning trajectory information and driving strategy information for the vehicle, and finally controls the vehicle based on the planning trajectory information and the driving strategy information.

[0070] In the disclosed embodiment, since the behavior trajectory information of the target object in the vehicle's environment and the vehicle's own trajectory information are first sent to the server, the server can generate guiding planning decision information for the vehicle based on the driving experience of the pre-trained large model (generating a pre-trained transformer). In this way, when the vehicle performs trajectory planning, it is not only based on the behavior trajectory information of the target object in the environment, but also combined with the guiding planning decision information fed back by the large model, thereby improving the safety of the generated planning trajectory information and the applicability of the driving strategy information, which is beneficial to improving the safety of autonomous driving and the user's driving experience.

[0071] To facilitate understanding of this embodiment, the execution entity of a vehicle control method disclosed in this embodiment is first described in detail. For example, the execution entity of the vehicle control method provided in this embodiment can be a vehicle, which can include various controllers, such as a vehicle controller, a body domain controller, a cockpit domain controller, an intelligent driving domain controller, etc., without specific limitation.

[0072] The controller may include a processor and a memory, the memory is used to store instructions, the processor calls the instructions from the memory and can implement the vehicle control method described in each embodiment of the present disclosure when executing the instructions.

[0073] In other embodiments, the vehicle control method may be executed by an onboard device or a chip within the onboard device, where the onboard device may be a vehicle computer or a device located on the vehicle. Specifically, the onboard device may include, for example, a terminal device or other processing device, such as a user terminal, a handheld device, a computing device, an onboard device, a wearable device, or the like.

[0074] In addition, the vehicle control method can also be implemented by a processor calling computer-readable instructions stored in a memory.

[0075] The vehicle control method provided by the embodiment of the present disclosure is described below with reference to the accompanying drawings. As shown in FIG1 , the vehicle control method includes the following steps S101 to S104:

[0076] S101, obtaining environmental perception data of the vehicle's environment, the vehicle's position data, and map data of the vehicle's environment, and determining behavior trajectory information of a target object in the vehicle's environment based on the environmental perception data, the vehicle's position data, and the map data.

[0077] For example, the vehicle is an autonomous vehicle with an autonomous driving function. The autonomous vehicle can rely on artificial intelligence, visual computing, radar, monitoring devices, and a global positioning system to work together to enable the vehicle's computer to automatically and safely operate the motor vehicle without any active human intervention.

[0078] Among them, the vehicle can include battery electric vehicles (BEV), hybrid electric vehicles (HEV) and fuel vehicles, etc. The specific vehicle type is not limited, as long as it has automatic driving function.

[0079] Specifically, environmental perception data of the vehicle's environment can be acquired through a perception component provided on the vehicle. The perception component may include a radar device and / or an image acquisition device. That is, in the embodiments of the present disclosure, the environmental perception data may include image data and / or radar data.

[0080] The radar device is used to collect point cloud data. In one example, the radar device can be a lidar, which emits a laser beam to detect target characteristics such as position and velocity. Each point in the point cloud data collected by the radar device contains three-dimensional coordinate information, and may also contain information such as the target, position, velocity, acceleration, reflection intensity, and echo count. In other examples, the radar device can also be a millimeter-wave radar, ultrasonic radar, etc.

[0081] The image acquisition device is used to collect image data of the vehicle's environment. In one example, the image acquisition device can be a camera. To ensure more accurate acquisition, the vehicle can be equipped with multiple image acquisition devices. The type of the image acquisition device is not specifically limited, for example, it can be a monocular camera or a binocular camera.

[0082] In addition, the vehicle's location data can be obtained through the global positioning system and the map data of the vehicle's environment can be obtained through the navigation system. Of course, in other embodiments, the location data and map data can also be obtained through other means, which are not specifically limited here.

[0083] After obtaining the environmental perception data of the vehicle's environment, the vehicle's position data and the map data of the vehicle's environment, the behavior trajectory information of the target object in the vehicle's environment can be determined based on the environmental perception data, the vehicle's position data and the map data.

[0084] Target objects include various dynamic and static obstacles in the vehicle's environment. Dynamic obstacles can include other moving motor vehicles, non-motor vehicles, and pedestrians; static obstacles include other stationary vehicles and various stationary objects on the road (such as flower beds, trees, and roadblocks).

[0085] In some embodiments, the target object's behavior trajectory information includes the target object's tracking trajectory information, the target object's predicted trajectory information, and the target object's predicted behavior information. The tracking trajectory information refers to the target object's trajectory information before the current moment, the predicted trajectory information refers to the target object's trajectory information within a preset time period after the current moment, and the predicted behavior information refers to the target object's predicted action after the current moment. For example, in the case where the target object is another vehicle, the predicted behavior can be the other vehicle's lane changing behavior or parking behavior, etc.

[0086] In the embodiment of the present disclosure, since the behavior trajectory information of the target object includes the tracking trajectory information that has already occurred, the predicted trajectory information that is about to occur, and the predicted behavior information that is about to occur, the behavior trajectory of the target object can be estimated and measured from multiple aspects, which is conducive to improving the accuracy of the trajectory behavior judgment of the target object.

[0087] In addition, the process of determining the behavior trajectory information of the target object based on the environmental perception data, the vehicle position data and the map data will be explained in detail later.

[0088] S102, sending the target object's behavior trajectory information and the vehicle's own trajectory information to a server, so that the server processes the target object's behavior trajectory information and the vehicle's own trajectory information based on a trained target model to generate planning decision information; the target model is a pre-trained transformer generated by self-supervised training based on multiple training sample data.

[0089] In this step, after the behavior trajectory information of the target object is determined on the vehicle side, the behavior trajectory information of the target object and the vehicle's own trajectory information can be sent to the server, so that the server can process (predict) the behavior trajectory information of the target object and the vehicle's own trajectory information based on the trained target model to generate planning decision information.

[0090] In the disclosed embodiment, the target model is a pre-trained transformer (also known as a large model) that is self-supervised and trained based on a variety of training sample data. Thus, upon receiving the target object's trajectory information and the vehicle's own trajectory information, the server can input the target object's trajectory information and the vehicle's own trajectory information into the pre-trained large model, causing the large model to output corresponding planning decision information.

[0091] Understandably, due to the large number of parameters in this large model, deploying it in the cloud can ensure its performance requirements. Furthermore, because this large model is trained through self-supervision based on multiple training data, it can output guiding planning and decision-making information based on the target object's behavioral trajectory information and the vehicle's own trajectory information. This planning and decision-making information is non-real-time data.

[0092] In other embodiments, the vehicle type information may also be sent to the server so that the server can combine the vehicle type factor in the process of generating planning decision information based on the target model, thereby improving the applicability of the planning decision information.

[0093] Among them, the server can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms.

[0094] In some embodiments, the planning decision information may include at least one of predicted trajectory information, driving instruction information, and driving style information for the vehicle. The trajectory length indicated by the predicted trajectory information is greater than a preset length, i.e., the predicted trajectory information represents the long trajectory information planned by the server for the vehicle. The preset length can be set based on actual needs but should be greater than the length of the planned trajectory information planned by the vehicle itself.

[0095] The driving instruction information is used to indicate the driving behavior of the vehicle. For example, the driving instruction information may include instruction information such as going straight, turning left, turning right, entering a ramp, and exiting a ramp.

[0096] The driving style information is used to indicate the smoothness of vehicle driving. For example, the driving style information may include smooth, aggressive, and courteous styles.

[0097] It should be noted that the training process of the target model will be explained in detail later.

[0098] S103: Receive the planning decision information fed back by the server, and generate planning trajectory information and driving strategy information for the vehicle according to the behavior trajectory information of the target object and the planning decision information.

[0099] For example, after receiving the planning decision information fed back by the server, planning trajectory information and driving strategy information for the vehicle can be generated based on the target object's behavioral trajectory information and the planning decision information. Specifically, the generation of planning trajectory information and driving strategy information can be achieved based on a deep learning model or decision tree.

[0100] Among them, the planned trajectory information is used to indicate at least one route information that the vehicle can travel, and the driving strategy information is used to reflect the driving behavior of the vehicle. Specifically, the driving strategy information may include driving speed information, acceleration information, vehicle gear information, lane change information, etc., without specific limitation.

[0101] In one example, as previously described, the planning decision information includes predicted trajectory information and driving instruction information for the vehicle. Therefore, when generating the planned trajectory information and driving strategy information for the vehicle based on the target object's behavioral trajectory information and the planning decision information, the following steps may be performed: generating the planned trajectory information and driving strategy information for the vehicle based on the target object's behavioral trajectory information, the predicted trajectory information, and the driving instruction information. This can improve the comfort and smoothness of the generated planned trajectory information and driving strategy information, thereby enhancing the driving experience of autonomous driving.

[0102] S104: Control the vehicle driving based on the planned trajectory information and the driving strategy information.

[0103] For example, after obtaining the planned trajectory information and driving strategy information for the vehicle, the vehicle can be controlled to travel according to the planned trajectory information based on the driving strategy information.

[0104] In some embodiments, the planning decision information also includes driving style information. Therefore, controlling the vehicle based on the planned trajectory information and the driving strategy information may include controlling the vehicle based on the planned trajectory information, the driving strategy information, and the driving style information. In this way, the driving style information provided by the large model is used as a guide during vehicle control, further improving vehicle control comfort.

[0105] Specifically, the vehicle can be controlled based on the planned trajectory, driving strategy, and driving style information from the large-scale model, taking into account optimal vehicle driving energy consumption. For example, the vehicle's throttle, brake, and steering wheel signals are calculated based on the planned trajectory to control the host vehicle to follow the specified trajectory, and feedback control is provided to respond to changes in vehicle dynamics and the environment.

[0106] Furthermore, during vehicle control, a hybrid vehicle model can be used to predict vehicle behavior and status, adjusting driving strategies accordingly. Driving style information can also be used to modify driving styles by adjusting key parameters (such as speed and acceleration constraints and following distance). For example, if the large model indicates an aggressive driving style, but control parameters during driving indicate that the vehicle's current speed and acceleration are low, the driving style can be modified to a courteous one.

[0107] The vehicle control method provided by the embodiment of the present disclosure first sends the behavior trajectory information of the target object in the vehicle's environment and the vehicle's own trajectory information to the server, so that the server can generate guiding planning decision information for the vehicle based on the driving experience of a pre-trained large model (generating a pre-trained transformer). In this way, when the vehicle performs trajectory planning, it is not only based on the behavior trajectory information of the target object in the environment, but also combined with the guiding planning decision information provided by the large model. This can improve the safety of the generated planning trajectory information and the applicability of the driving strategy information, thereby improving the safety of autonomous driving and the user's driving experience.

[0108] The following describes in detail a process of determining the behavior trajectory information of a target object in the environment where the vehicle is located based on the environment perception data, the vehicle position data, and the map data.

[0109] In some embodiments, the environmental perception data includes image data. Therefore, when determining the behavior trajectory information of a target object in the environment in which the vehicle is located based on the environmental perception data, the vehicle's position data, and the map data, the following (a) to (c) may be included:

[0110] (a) performing feature extraction on the image data to obtain bird's-eye view features of the image data, and determining bird's-eye view features of the map data based on the bird's-eye view features of the image data;

[0111] (b) fusing the bird's-eye view feature of the image data with the bird's-eye view feature of the map data based on the position data of the vehicle to obtain a target fusion feature;

[0112] (c) Determining the behavior trajectory information of the target object based on the target fusion features.

[0113] The image data may be image data collected based on a single image acquisition device (such as surround view image data), or image data obtained by separately collecting images based on multiple image acquisition devices with different perspectives and fusing the collected images.

[0114] Exemplarily, as shown in FIG2 , perspective view (PV) features of image data can be extracted based on a preset feature extraction network to obtain perspective view features, and the perspective view features can be converted to obtain bird's eye view (BEV) features of the image data.

[0115] The feature extraction network may include, but is not limited to, a RegNet neural network or a RestNet neural network. Furthermore, a Transformer network may be used to transform the perspective feature. In some embodiments, image data from different cameras may be dedistorted based on their intrinsic parameters before feature extraction.

[0116] After obtaining the bird's-eye view features of the image data, the bird's-eye view features of the map data can be determined based on the bird's-eye view features of the image data. Specifically, the map data can be sampled according to the physical dimensions of the bird's-eye view features of the image data, that is, the physical distance represented by each pixel is consistent with the bird's-eye view features of the image data. Then, feature extraction is performed on the map data based on a convolutional neural network or a self-attention network to obtain the bird's-eye view features of the map data.

[0117] It should be noted that map data can be determined based on the vehicle's absolute position data and a system map. The system map can be a high-precision map or a standard map, which is not specifically limited here. The vehicle's absolute position data can be determined based on the vehicle's relative position data and the vehicle's position data. For example, the vehicle's position data and relative position data can be fused through optimization or filtering methods to obtain the vehicle's absolute position data.

[0118] For example, the vehicle's wheel speed data and inertial measurement unit (IMU) inertial navigation data can be obtained, and then the vehicle's relative posture data can be determined based on the vehicle's wheel speed data and inertial navigation data. Then, the vehicle's absolute posture data can be determined based on the vehicle's position data and the vehicle's relative posture data.

[0119] In addition, the range corresponding to the map data is larger than the range corresponding to the image data. For example, the range corresponding to the map data may be twice the range corresponding to the image data, which is not specifically limited.

[0120] It is understandable that due to the continuous movement of the vehicle itself, the BEV features at two moments are not spatially aligned. Therefore, it is necessary to obtain the vehicle's own rotation angle and offset information based on the relative posture change of the vehicle, so as to spatially align the BEV features of the previous moment with the BEV features of the current moment.

[0121] Therefore, in some embodiments, when the bird's-eye view feature of the image data is fused with the bird's-eye view feature of the map data based on the vehicle's position data to obtain the target fusion feature, the following (1) to (2) may be included:

[0122] (1) fusing the bird's-eye view feature of the image data with the bird's-eye view feature of the map data at the corresponding moment to obtain a single-frame fused bird's-eye view feature;

[0123] (2) Based on the position data of the vehicle, the relative posture data of the vehicle at adjacent moments are determined, and based on the relative posture data, the multi-frame fused bird's-eye view features at the current moment and before the current moment are aligned and fused to obtain the target fusion feature.

[0124] The specific number of multiple frames can be determined according to actual needs and is not limited here.

[0125] It is understood that in other embodiments, the BEV features of multiple frames of image data may be first spatially and temporally aligned to obtain target BEV features of the image data. The map is then sampled and feature extracted based on the target BEV features to obtain target BEV features of the map data. The target BEV features of the image data are then fused with the target BEV features of the map data to obtain the target fused features. For example, the multi-scale surround view image features extracted by the backbone network and the Neck network from the PV perspective can be queried using a spatial cross-attention module to generate BEV fused features in the BEV space.

[0126] After obtaining the target fusion features, the target object's behavior trajectory information can be determined based on the target fusion features. Specifically, the target object's behavior trajectory information can be predicted based on machine learning methods or deep learning methods. For example, the target object's behavior information (such as lane changing, parking, etc.) can be predicted based on a behavior classification network (convolutional neural network, recurrent neural network, etc.).

[0127] It should be noted that while the vehicle is moving, the objects around the vehicle are also moving within a certain range. The alignment of objects around the vehicle can be corrected by learning based on the network's own time-domain self-attention module.

[0128] As shown in FIG3 , in some embodiments, after obtaining the target fusion feature, the method further includes: determining a plurality of perception information based on the target fusion feature, where the plurality of perception information includes at least two of the following:

[0129] Three-dimensional object perception information (BEV 3D moving objects), three-dimensional road structure perception information (BEV 3D road structure), occupied space grid perception information (Occupancy & Flow), traffic light perception information (traffic light recognition), three-dimensional map perception information (online mapping) and fusion positioning perception information (Model Based Fusion Positioning).

[0130] In the embodiment of the present disclosure, after obtaining the target fusion feature, multiple perception information can be determined based on the target fusion feature based on the task network, and the multiple perception information can be output separately. In this way, more references can be provided for subsequent trajectory information planning, which is conducive to improving the effectiveness and safety of trajectory planning.

[0131] Therefore, in some embodiments, when generating the planned trajectory information and driving strategy information for the vehicle based on the target object's behavioral trajectory information and the planning decision information, the process may include: generating the planned trajectory information and driving strategy information for the vehicle based on the target object's behavioral trajectory information, the traffic light perception information, the three-dimensional map perception information, and the planning decision information. This can further improve the effectiveness and comfort of the vehicle's planned trajectory information and driving strategy information, thereby further enhancing the autonomous driving experience.

[0132] 4 is a flowchart of a vehicle control method according to another embodiment of the present disclosure. In this embodiment, the vehicle control method is applied to the aforementioned server, that is, the execution subject of the vehicle control method is the server. The vehicle control method includes the following steps S401 to S403:

[0133] S401, receiving vehicle trajectory information for the target vehicle and behavior trajectory information of a target object in the environment of the target vehicle sent by the target vehicle; the behavior trajectory information of the target object is determined based on environmental perception data of the environment in which the vehicle is located, the position data of the vehicle, and the map data of the environment in which the vehicle is located.

[0134] S402 , processing the target object's behavior trajectory information and the vehicle's trajectory information based on a trained target model to generate planning decision information; the target model is a pre-trained transformer, and is obtained through self-supervised training based on a variety of training sample data.

[0135] S403, sending the planning decision information to the target vehicle, wherein the planning decision information is used to instruct the target vehicle to generate planning trajectory information and driving strategy information for the vehicle based on the behavior trajectory information of the target object and the planning decision information.

[0136] The contents of steps S401 to S403 may refer to the contents of steps S101 to S104 described above, and will not be repeated here.

[0137] In one possible implementation, the target model is trained by the following steps (I) to (II):

[0138] (1) obtaining a basic network to be trained, wherein the model parameters of the basic network are greater than a preset scale;

[0139] (II) obtaining training sample data, and performing self-supervised training on the base network based on the training sample data to obtain the trained target model; the training sample data includes at least two of the following:

[0140] Vehicle trajectory sample data, driving status sample data, weather condition sample data during vehicle driving, traffic condition sample data during vehicle driving, navigation route sample data corresponding to vehicle driving, road structure status sample data perceived by the vehicle, trajectory sample data of dynamic and static obstacles, and Internet driving video sample data.

[0141] Specifically, as shown in Figure 5, during model training, multimodal feature extraction is first achieved through self-supervised training of a basic network. This network training utilizes large-scale vehicle trajectory sample data, driving state sample data (gear shifting, wheel speed, steering wheel angle, acceleration and deceleration, etc.), road weather condition sample data during driving (season, sunny, cloudy, rainy, snowy, etc., morning, noon, and evening time, etc.), road traffic condition sample data (smooth, morning and evening rush hour, congestion, etc.), planned path and navigation information sample data from the corresponding map of the vehicle driving, sample data on the surrounding road structure state perceived by the vehicle, and sample data on the trajectory of dynamic and static obstacles. In addition, relevant driving videos and related annotations and labels on the internet can also be used to train the basic network.

[0142] In the disclosed embodiment, since the target model is a large model trained based on a variety of training data, the trained target model is not only more applicable, but also improves the prediction accuracy of the model.

[0143] In some embodiments, after basic training of the network, it is also possible to use the vehicle trajectory and driving decision data of normal human driving on the basis of the pre-trained basic feature model to perform non-real-time, multi-group decision instructions and multiple long planning trajectories based on imitation learning.

[0144] In addition, generalized testing can be used to find corner cases of decision instructions and planning trajectories. At the same time, rules can be used to clarify the safe driving framework to constrain the planning trajectory of the large model, and decision instructions and planning trajectories that deviate from the safe driving framework can be included in extreme cases. Then, using extreme case samples, deep reinforcement learning can be used to improve the generation of decision instructions and planning trajectories. In this way, the target model obtained can provide planning decision information with stronger guiding significance.

[0145] Those skilled in the art will understand that in the above-mentioned method of the specific implementation method, the writing order of each step does not mean a strict execution order and does not constitute any limitation on the implementation process. The specific execution order of each step should be determined by its function and possible internal logic.

[0146] FIG6 is a functional block diagram of a vehicle control device according to an embodiment of the present disclosure. The vehicle control device 600 includes:

[0147] An information determination module 601 is configured to obtain environmental perception data of an environment in which a vehicle is located, the vehicle's position data, and map data of the environment in which the vehicle is located, and determine, based on the environmental perception data, the vehicle's position data, and the map data, behavior trajectory information of a target object in the environment in which the vehicle is located;

[0148] A first sending module 602 is configured to send the target object's behavior trajectory information and the vehicle's own trajectory information to a server, so that the server processes the target object's behavior trajectory information and the vehicle's own trajectory information based on a trained target model to generate planning decision information; the target model is a pre-trained transformer obtained through self-supervised training based on a plurality of training sample data;

[0149] An information generation module 603 is configured to receive the planning decision information fed back by the server and generate planning trajectory information and driving strategy information for the vehicle based on the target object's behavior trajectory information and the planning decision information;

[0150] The vehicle control module 604 is configured to control the vehicle based on the planned trajectory information and the driving strategy information.

[0151] In one possible implementation, the planning decision information includes predicted trajectory information and driving instruction information for the vehicle, where the trajectory length indicated by the predicted trajectory information is greater than a preset length, and the driving instruction information is used to indicate the driving behavior of the vehicle; the information generation module 603 is specifically configured to:

[0152] According to the behavior trajectory information of the target object, the predicted trajectory information and the driving instruction information, planned trajectory information and driving strategy information for the vehicle are generated.

[0153] In a possible implementation, the planning decision information further includes driving style information, where the driving style information is used to indicate a degree of smoothness of vehicle driving; the vehicle control module 604 is specifically configured to:

[0154] The vehicle is controlled to travel based on the planned trajectory information, the driving strategy information, and the driving style information.

[0155] In a possible implementation, the behavior trajectory information of the target object includes tracking trajectory information of the target object, predicted trajectory information of the target object, and predicted behavior information of the target object.

[0156] In a possible implementation, the information determination module 601 is specifically configured to:

[0157] performing feature extraction on the image data to obtain bird's-eye view features of the image data, and determining bird's-eye view features of the map data based on the bird's-eye view features of the image data;

[0158] Based on the position data of the vehicle, fusing the bird's-eye view feature of the image data with the bird's-eye view feature of the map data to obtain a target fusion feature;

[0159] Based on the target fusion features, behavior trajectory information of the target object is determined.

[0160] In a possible implementation, the information determination module 601 is specifically configured to:

[0161] fusing the bird's-eye view feature of the image data with the bird's-eye view feature of the map data at a corresponding moment to obtain a single-frame fused bird's-eye view feature;

[0162] Based on the position data of the vehicle, the relative posture data of the vehicle at adjacent moments are determined, and based on the relative posture data, the multi-frame fused bird's-eye view features at the current moment and before the current moment are aligned and fused to obtain the target fusion feature.

[0163] In a possible implementation, the information determination module 601 is further configured to:

[0164] Determine a plurality of perception information based on the target fusion feature, wherein the plurality of perception information includes at least two of the following:

[0165] Three-dimensional object perception information, three-dimensional road structure perception information, occupied space grid perception information, traffic light perception information, three-dimensional map perception information and fused positioning perception information.

[0166] In a possible implementation, the information generation module 603 is specifically configured to:

[0167] Based on the behavior trajectory information of the target object, the traffic light perception information, the three-dimensional map perception information and the planning decision information, planning trajectory information and driving strategy information for the vehicle are generated.

[0168] FIG7 is a functional block diagram of another vehicle control device according to an embodiment of the present disclosure. The vehicle control device 700 includes:

[0169] An information receiving module 701 is configured to receive vehicle trajectory information of a target vehicle and behavior trajectory information of a target object in the environment of the target vehicle, the behavior trajectory information of the target object being determined based on environmental perception data of the vehicle's environment, the vehicle's location data, and map data of the vehicle's environment.

[0170] An information processing module 702 is configured to process the target object's behavior trajectory information and the vehicle's trajectory information based on a trained target model to generate planning decision information; the target model is a pre-trained transformer obtained through self-supervised training based on a variety of training sample data;

[0171] The second sending module 703 is used to send the planning decision information to the target vehicle, and the planning decision information is used to instruct the target vehicle to generate planning trajectory information and driving strategy information for the vehicle based on the behavior trajectory information of the target object and the planning decision information.

[0172] In a possible implementation, referring to FIG8 , the vehicle control device 700 further includes a model training module 704 , and the model training module 704 is configured to:

[0173] Obtaining a basic network to be trained, wherein model parameters of the basic network are greater than a preset scale;

[0174] Acquire training sample data, and perform self-supervised training on the basic network based on the training sample data to obtain the trained target model; the training sample data includes at least two of the following:

[0175] Vehicle trajectory sample data, driving status sample data, weather condition sample data during vehicle driving, traffic condition sample data during vehicle driving, navigation route sample data corresponding to vehicle driving, road structure status sample data perceived by the vehicle, trajectory sample data of dynamic and static obstacles, and Internet driving video sample data.

[0176] Based on the same technical concept, the present disclosure also provides an in-vehicle device. Referring to FIG9 , a schematic diagram of the structure of an in-vehicle device 900 provided by the present disclosure is shown, which includes a processor 901, a memory 902, and a bus 903. The memory 902 is used to store execution instructions.

[0177] In the disclosed embodiment, the memory 902 is specifically used to store application code for executing the solution of the present application, and the execution is controlled by the processor 901. That is, when the vehicle-mounted device 900 is running, the processor 901 communicates with the memory 902 via the bus 903, so that the processor 901 executes the application code stored in the memory 902, thereby performing the method described in any of the aforementioned embodiments.

[0178] Among them, the memory 902 can be, but is not limited to, random access memory (RAM), read only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), etc.

[0179] The processor 901 may be an integrated circuit chip with signal processing capabilities. The above-mentioned processor may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it may also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. The various methods, steps and logic block diagrams disclosed in the embodiments of the present invention can be implemented or executed. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0180] It should be understood that the structure illustrated in the embodiments of the present application does not constitute a specific limitation on the vehicle-mounted device 900. In other embodiments of the present application, the vehicle-mounted device 900 may include more or fewer components than illustrated, or may combine or separate certain components, or may have different component arrangements. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.

[0181] In addition, it should be noted that the memory and processor included in the vehicle controller are similar to the processor 901 and memory 902 included in the vehicle-mounted device 900, and will not be repeated here.

[0182] An embodiment of the present disclosure further provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the vehicle control method in the above method embodiment are executed.

[0183] An embodiment of the present disclosure also provides a computer program product, including a computer program / instruction. When the computer program / instruction processor is executed, the vehicle control method provided in each embodiment of the present disclosure is implemented. For details, please refer to the above method embodiments, which will not be repeated here.

[0184] The computer program product may be implemented in hardware, software, or a combination thereof. In one embodiment, the computer program product is implemented as a computer storage medium. In another embodiment, the computer program product is implemented as a software product, such as a software development kit (SDK).

[0185] The methods in the embodiments of the present disclosure may be implemented in whole or in part through software, hardware, firmware, or any combination thereof. When implemented using software, they may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer programs or instructions. When the computer program or instructions are downloaded and executed on a computer, the processes or functions described herein are executed in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, a network device, a user device, a core network device, an OAM, or other programmable device.

[0186] The computer program or instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer program or instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via a wired or wireless method. The computer-readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server or data center that integrates one or more available media. The available medium may be a magnetic medium, such as a floppy disk, a hard disk, or a magnetic tape; an optical medium, such as a digital video disk; or a semiconductor medium, such as a solid-state drive. The computer-readable storage medium may be a volatile or non-volatile storage medium, or may include both volatile and non-volatile types of storage media.

[0187] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems and devices described above can refer to the corresponding processes in the aforementioned method embodiments, and will not be repeated here. In the several embodiments provided in the present disclosure, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0188] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of the solution of this embodiment according to actual needs.

[0189] In addition, each functional unit in each embodiment of the present disclosure may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0190] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium that is executable by a processor. Based on this understanding, the technical solution of the present disclosure, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present disclosure. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a ROM, a RAM, a magnetic disk, or an optical disk.

[0191] Finally, it should be noted that the above-described embodiments are only specific implementation methods of the present disclosure, which are used to illustrate the technical solutions of the present disclosure, rather than to limit them. The scope of protection of the present disclosure is not limited thereto. Although the present disclosure has been described in detail with reference to the above-described embodiments, those skilled in the art should understand that any person skilled in the art can modify or easily conceive of changes to the technical solutions described in the above-described embodiments within the technical scope disclosed in the present disclosure, or replace some of the technical features therein with equivalents. Such modifications, changes, or replacements do not deviate from the spirit and scope of the technical solutions of the embodiments of the present disclosure, and should be included in the scope of protection of the present disclosure. Therefore, the scope of protection of the present disclosure shall be subject to the scope of protection of the claims.

Claims

1. A vehicle control method, characterized in that: include: Acquiring environmental perception data of an environment in which a vehicle is located, position data of the vehicle, and map data of the environment in which the vehicle is located, and determining behavior trajectory information of a target object in the environment in which the vehicle is located based on the environmental perception data, the position data of the vehicle, and the map data; The target object's behavior trajectory information and the vehicle's own trajectory information are sent to a server, so that the server processes the target object's behavior trajectory information and the vehicle's own trajectory information based on a trained target model to generate planning decision information; the target model is a pre-trained transformer obtained through self-supervised training based on a plurality of training sample data; receiving the planning decision information fed back by the server, and generating planning trajectory information and driving strategy information for the vehicle based on the behavior trajectory information of the target object and the planning decision information; The vehicle is controlled based on the planned trajectory information and the driving strategy information.

2. The method according to claim 1, characterized in that The planning decision information includes predicted trajectory information and driving instruction information for the vehicle, wherein the trajectory length indicated by the predicted trajectory information is greater than a preset length, and the driving instruction information is used to instruct the driving behavior of the vehicle; The generating of planning trajectory information and driving strategy information for the vehicle based on the behavior trajectory information of the target object and the planning decision information includes: According to the behavior trajectory information of the target object, the predicted trajectory information and the driving instruction information, planned trajectory information and driving strategy information for the vehicle are generated.

3. The method according to claim 1, characterized in that The planning decision information further includes driving style information, wherein the driving style information is used to indicate the smoothness of vehicle driving; The controlling the vehicle driving based on the planned trajectory information and the driving strategy information includes: The vehicle is controlled to travel based on the planned trajectory information, the driving strategy information, and the driving style information.

4. The method according to claim 1, wherein The behavior trajectory information of the target object includes tracking trajectory information of the target object, predicted trajectory information of the target object, and predicted behavior information of the target object.

5. The method according to claim 4, characterized in that The environmental perception data includes image data, and determining the behavior trajectory information of the target object in the environment where the vehicle is located based on the environmental perception data, the position data of the vehicle, and the map data includes: performing feature extraction on the image data to obtain bird's-eye view features of the image data, and determining bird's-eye view features of the map data based on the bird's-eye view features of the image data; Based on the position data of the vehicle, fusing the bird's-eye view feature of the image data with the bird's-eye view feature of the map data to obtain a target fusion feature; Based on the target fusion features, behavior trajectory information of the target object is determined.

6. The method according to claim 5, characterized in that The step of fusing the bird's-eye view feature of the image data with the bird's-eye view feature of the map data based on the vehicle position data to obtain a target fusion feature includes: fusing the bird's-eye view feature of the image data with the bird's-eye view feature of the map data at a corresponding moment to obtain a single-frame fused bird's-eye view feature; Based on the position data of the vehicle, the relative posture data of the vehicle at adjacent moments are determined, and based on the relative posture data, the multi-frame fused bird's-eye view features at the current moment and before the current moment are aligned and fused to obtain the target fusion feature.

7. The method according to claim 5, characterized in that After obtaining the target fusion feature, the method further includes: Determine a plurality of perception information based on the target fusion feature, wherein the plurality of perception information includes at least two of the following: Three-dimensional object perception information, three-dimensional road structure perception information, occupied space grid perception information, traffic light perception information, three-dimensional map perception information and fused positioning perception information.

8. The method according to claim 7, characterized in that The generating of planning trajectory information and driving strategy information for the vehicle based on the behavior trajectory information of the target object and the planning decision information includes: Based on the behavior trajectory information of the target object, the traffic light perception information, the three-dimensional map perception information and the planning decision information, planning trajectory information and driving strategy information for the vehicle are generated.

9. A vehicle control method, characterized in that: include: Receiving vehicle trajectory information for the target vehicle and behavior trajectory information of target objects in the environment of the target vehicle sent by the target vehicle; The behavior trajectory information of the target object is determined based on environmental perception data of the environment in which the vehicle is located, the position data of the vehicle, and map data of the environment in which the vehicle is located; Processing the target object's behavior trajectory information and the vehicle's trajectory information based on the trained target model to generate planning decision information; The target model is a pre-trained transformer, which is obtained through self-supervised training based on a variety of training sample data; The planning decision information is sent to the target vehicle, where the planning decision information is used to instruct the target vehicle to generate planning trajectory information and driving strategy information for the vehicle based on the behavior trajectory information of the target object and the planning decision information.

10. The method according to claim 9, characterized in that: The target model is trained by the following steps: Obtaining a basic network to be trained, wherein model parameters of the basic network are greater than a preset scale; Acquire training sample data, and perform self-supervised training on the basic network based on the training sample data to obtain the trained target model; the training sample data includes at least two of the following: Vehicle trajectory sample data, driving status sample data, weather condition sample data during vehicle driving, traffic condition sample data during vehicle driving, navigation route sample data corresponding to vehicle driving, road structure status sample data perceived by the vehicle, trajectory sample data of dynamic and static obstacles, and Internet driving video sample data.

11. A vehicle control device, characterized in that: include: an information determination module, configured to obtain environmental perception data of an environment in which a vehicle is located, position data of the vehicle, and map data of the environment in which the vehicle is located, and determine, based on the environmental perception data, position data of the vehicle, and map data, behavior trajectory information of a target object in the environment in which the vehicle is located; a first sending module configured to send the target object's behavior trajectory information and the vehicle's own trajectory information to a server, so that the server processes the target object's behavior trajectory information and the vehicle's own trajectory information based on a trained target model to generate planning decision information; the target model is a pre-trained transformer obtained through self-supervised training based on a plurality of training sample data; an information generation module, configured to receive the planning decision information fed back by the server, and generate planning trajectory information and driving strategy information for the vehicle based on the behavior trajectory information of the target object and the planning decision information; A vehicle control module is used to control the vehicle driving based on the planned trajectory information and the driving strategy information.

12. A vehicle control device, characterized in that: include: An information receiving module is used to receive the vehicle trajectory information of the target vehicle and the behavior trajectory information of the target object in the environment of the target vehicle sent by the target vehicle; The behavior trajectory information of the target object is determined based on environmental perception data of the environment in which the vehicle is located, the position data of the vehicle, and map data of the environment in which the vehicle is located; An information processing module, configured to process the target object's behavior trajectory information and the vehicle's trajectory information based on the trained target model to generate planning decision information; The target model is a pre-trained transformer, which is obtained through self-supervised training based on a variety of training sample data; The second sending module is used to send the planning decision information to the target vehicle, and the planning decision information is used to instruct the target vehicle to generate planning trajectory information and driving strategy information for the vehicle based on the behavior trajectory information of the target object and the planning decision information.

13. A vehicle-mounted device, characterized in that: include: A processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor. When the vehicle-mounted device is running, the processor and the memory communicate via the bus. When the machine-readable instructions are executed by the processor, the vehicle control method according to any one of claims 1 to 8 is executed.

14. A vehicle, characterized in that: A controller is included, the controller comprising: a memory configured to store instructions; and A processor is configured to call the instructions from the memory and implement the vehicle control method according to any one of claims 1 to 8 when executing the instructions.

15. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, which, when executed by a processor, executes the vehicle control method according to any one of claims 1 to 10.

Images