Traffic scene construction method, electronic equipment, storage medium and program product

Abstract

The invention provides a traffic scene construction method, electronic equipment, a storage medium and a program product, and relates to the technical field of automatic driving. The method comprises the steps that motion data and traffic scene data of a target vehicle in a target traffic scene are acquired, the motion data are determined based on first motion track data and / or attitude of the target vehicle in the target traffic scene, and the traffic scene data are acquired by the target vehicle; according to the traffic scene data, scene motion information corresponding to each first traffic participant of the target vehicle is determined, and the scene motion information is determined based on second motion track data and / or postures of the corresponding first traffic participants in the target traffic scene; the first traffic participation objects comprise objects except the target vehicle in the target traffic scene; and constructing a target traffic scene at least according to the motion data and the scene motion information corresponding to each associated traffic participation object. According to the method, the coverage rate of the constructed target traffic scene is improved.

Description

Technical Field

[0001] The present application relates to the field of autonomous driving technology, and in particular to a traffic scene construction method, electronic equipment, storage medium and program product. Background Art

[0002] With the rapid development of autonomous driving technology, building accurate and comprehensive traffic scenarios has become one of the key links in realizing a safe and efficient autonomous driving system. Therefore, there is an urgent need for a construction method that can provide accurate and comprehensive traffic scenarios.

[0003] At present, traffic data in the current fixed scene is usually collected through roadside units, such as roadside cameras or low-beam laser radars, and target monitoring and information extraction are further performed on the collected traffic data to construct traffic scenes. However, the traffic scenes constructed by this collection method have a limited range and limited detection targets, and the scene coverage is low. Summary of the invention

[0004] The present application provides a traffic scene construction method, electronic equipment, storage medium and program product to improve scene coverage.

[0005] In a first aspect, the present application provides a traffic scene construction method, comprising:

[0006] Acquire motion data and traffic scene data of the target vehicle in the target traffic scene, wherein the motion data is determined based on first motion trajectory data and / or posture of the target vehicle in the target traffic scene, and the traffic scene data is collected by the target vehicle;

[0007] Determine, according to the traffic scene data, scene motion information corresponding to each first traffic participant of the target vehicle, wherein the scene motion information is determined based on second motion trajectory data and / or posture of the corresponding first traffic participant in the target traffic scene; the first traffic participant object includes objects other than the target vehicle in the target traffic scene;

[0008] A target traffic scene is constructed based on at least the motion data and scene motion information corresponding to each associated traffic participant object.

[0009] In a possible implementation, the traffic scene construction method further includes:

[0010] Determine the traffic scene that the target vehicle passes through during the target period as the target traffic scene;

[0011] and / or,

[0012] determining a preset traffic scene as a target traffic scene;

[0013] and / or,

[0014] The traffic scene data includes at least one of images, videos, point clouds, and trajectories.

[0015] In a possible implementation, determining scene motion information corresponding to each first traffic participant of the target vehicle according to the traffic scene data includes:

[0016] Performing target recognition on the traffic scene data, and determining each recognized target as a first traffic participant;

[0017] Identifying current motion information of each first traffic participant in the target traffic scene based on the traffic scene data;

[0018] Determine the current motion information of each first traffic participant in the target traffic scene as the scene motion information corresponding to each first traffic participant; or

[0019] Based on the current motion information of each first traffic participant in the target traffic scene, the inferred motion information of each first traffic participant in the associated scene of the target traffic scene is estimated as the scene motion information corresponding to each first traffic participant; or

[0020] The inferred motion information of each first traffic participant in the associated scene of the target traffic scene and the current motion information of each first traffic participant in the target traffic scene are determined as the scene motion information corresponding to each first traffic participant; wherein

[0021] The associated scene is a historical scene before each first traffic participant enters the target traffic scene, and / or the associated scene is a future scene after each first traffic participant exits the target traffic scene.

[0022] In a possible implementation, constructing a target traffic scene at least according to the motion data and the scene motion information corresponding to each associated traffic participant object includes:

[0023] Based on the electronic map, correcting the motion data to obtain first motion data;

[0024] Based on the electronic map, correcting the scene motion information to obtain first scene motion information;

[0025] A target traffic scene is constructed according to the first motion data and the first scene motion information.

[0026] In a possible implementation, based on the electronic map, correcting the motion data to obtain the first motion data includes:

[0027] Acquire the initial position coordinates of the target vehicle based on the Global Navigation Satellite System (GNSS), and acquire the scene image collected by the target vehicle under the initial position coordinates;

[0028] Extract road features from scene images;

[0029] Determine a feature difference matrix based on the road elements in the scene image and the road elements in the electronic map;

[0030] The motion data is corrected according to the feature difference matrix to obtain first motion data.

[0031] In a possible implementation, correcting the motion data according to the feature difference matrix to obtain the first motion data includes:

[0032] Determine the position corresponding to any time within the set time period, and perform the following processing on the first motion trajectory data associated with the position: take the corrected first initial motion data as the initial value, and correct the second initial motion data of the target vehicle according to the feature difference matrix, the first initial motion data is the motion data associated with the position corresponding to the current time, and the second initial motion data is the motion data associated with the position corresponding to the next time;

[0033] The first motion data is obtained according to the motion data corrected at all times.

[0034] In a possible implementation, it further includes:

[0035] Determine the target lane of the target vehicle in the electronic map according to the initial position coordinates of the target vehicle based on GNSS;

[0036] Correct the motion data according to the target lane.

[0037] In a possible implementation, constructing a target traffic scene at least according to the motion data and the scene motion information corresponding to each associated traffic participant object includes:

[0038] Obtain the space occupied by the target vehicle in the target traffic scene;

[0039] Construct the target traffic scene based on motion data, scene motion information and occupied space.

[0040] In a second aspect, the present application provides a traffic scene construction device, comprising:

[0041] An acquisition module, configured to acquire the motion data and traffic scene data of a target vehicle in a target traffic scene, where the motion data is determined based on the first motion trajectory data and / or attitude of the target vehicle in the target traffic scene, and the traffic scene data is collected by the target vehicle;

[0042] A determination module, configured to determine the scene motion information corresponding to each first traffic participant of the target vehicle according to the traffic scene data, where the scene motion information is determined based on the second motion trajectory data and / or attitude of the corresponding first traffic participant in the target traffic scene; the first traffic participants include the objects other than the target vehicle in the target traffic scene;

[0043] A construction module, configured to construct the target traffic scene at least according to the motion data and the scene motion information corresponding to each associated traffic participant;

[0044] In a possible implementation manner, the determination module is specifically configured to:

[0045] Determine the traffic scene passed by the target vehicle during the target time period as the target traffic scene;

[0046] And / or,

[0047] Determine a preset traffic scene as the target traffic scene;

[0048] And / or,

[0049] The traffic scene data includes at least one of an image, a video, a point cloud, and a trajectory.

[0050] In a possible implementation manner, the determination module is specifically configured to:

[0051] Perform target recognition on the traffic scene data, and respectively determine the recognized targets as the first traffic participants;

[0052] Recognize the current motion information of each first traffic participant in the target traffic scene based on the traffic scene data;

[0053] Determine the current motion information of each first traffic participant in the target traffic scene as the scene motion information corresponding to each first traffic participant; or

[0054] Estimate the speculative motion information of each first traffic participant in the associated scene of the target traffic scene based on the current motion information of each first traffic participant in the target traffic scene as the scene motion information corresponding to each first traffic participant; or

[0055] Determine the speculative motion information of each first traffic participant in the associated scenario of the target traffic scenario and the current motion information of each first traffic participant in the target traffic scenario as the scenario motion information corresponding to each first traffic participant; where

[0056] The associated scenario is the historical scenario before each first traffic participant enters the target traffic scenario, and / or, the associated scenario is the future scenario after each first traffic participant exits the target traffic scenario.

[0057] In a possible implementation manner, the construction module is specifically configured to:

[0058] Based on the electronic map, correct the motion data to obtain the first motion data;

[0059] Based on the electronic map, correct the scenario motion information to obtain the first scenario motion information;

[0060] Construct the target traffic scenario according to the first motion data and the first scenario motion information.

[0061] In a possible implementation manner, the traffic scenario construction device further includes a processing module, and the processing module is specifically configured to:

[0062] Obtain the initial position coordinates of the target vehicle based on GNSS, and obtain the scene image collected by the target vehicle at the initial position coordinates;

[0063] Extract the road elements in the scene image;

[0064] Determine the feature difference matrix according to the road elements in the scene image and the road elements in the electronic map;

[0065] Correct the motion data according to the feature difference matrix to obtain the first motion data.

[0066] In a possible implementation manner, the processing module is specifically configured to:

[0067] Determine the position corresponding to any moment within a set time period, and perform the following processing on the motion data associated with the position: taking the corrected first initial motion data as the initial value, according to the feature difference matrix, correct the second initial motion data of the target vehicle, the first initial motion data is the motion data associated with the position corresponding to the current moment, and the second initial motion data is the motion data associated with the position corresponding to the next moment;

[0068] Obtain the first motion data according to the motion data corrected at all moments.

[0069] In a possible implementation manner, the processing module is specifically configured to:

[0070] Determine the target lane of the target vehicle in the electronic map according to the initial position coordinates of the target vehicle based on GNSS;

[0071] Correct the motion data according to the target lane.

[0072] In a possible implementation, the building block is further used to:

[0073] Obtain the space occupied by the target vehicle in the target traffic scene;

[0074] Construct the target traffic scene based on motion data, scene motion information and occupied space.

[0075] In a third aspect, the present application provides a model determination method, comprising:

[0076] Acquire scene data corresponding to a target traffic scene, where the target traffic scene is constructed through the first aspect and / or various possible implementation methods of the first aspect;

[0077] The scenario data is used to train or optimize the self-driving vehicle prediction planning model, which is used for vehicle path planning.

[0078] In a fourth aspect, the present application provides a model determination device, comprising:

[0079] An acquisition module, used to acquire scene data corresponding to a target traffic scene, where the target traffic scene is constructed through the first aspect and / or various possible implementations of the first aspect;

[0080] The processing module is used to use scene data to train or optimize the self-driving vehicle prediction planning model, which is used for vehicle path planning.

[0081] In a fifth aspect, the present application provides a path planning method, which is applied to a vehicle. The path planning method includes:

[0082] Get scene data;

[0083] The scene data is input into the self-vehicle prediction planning model to obtain the vehicle's path planning, and the self-vehicle prediction planning model is determined by the third aspect and / or various possible implementation methods of the third aspect.

[0084] In a sixth aspect, the present application provides a path planning device, which is applied to a vehicle, and the path planning device includes:

[0085] Acquisition module, used to acquire scene data;

[0086] A planning module is used to input scene data into the self-vehicle prediction planning model to obtain the vehicle's path planning. The self-vehicle prediction planning model is determined by the third aspect and / or various possible implementation methods of the third aspect.

[0087] In a seventh aspect, the present application provides an electronic device, including: a memory, a processor;

[0088] Memory stores computer-executable instructions;

[0089] The processor executes the computer-executable instructions stored in the memory, so that the processor executes the first aspect and / or the third aspect and / or the fifth aspect and / or various possible implementations of the first aspect and / or various possible implementations of the third aspect and / or various possible implementations of the fifth aspect as described above.

[0090] In an eighth aspect, the present application provides a computer-readable storage medium, in which computer execution instructions are stored. When the computer execution instructions are executed, they are used to implement the first aspect and / or the third aspect and / or the fifth aspect and / or various possible implementation methods of the first aspect and / or various possible implementation methods of the third aspect and / or various possible implementation methods of the fifth aspect as above.

[0091] In a ninth aspect, the present application provides a computer program product, comprising a computer program, which, when executed, implements the first aspect and / or the third aspect and / or the fifth aspect and / or various possible implementations of the first aspect and / or various possible implementations of the third aspect and / or various possible implementations of the fifth aspect.

[0092] The traffic scene construction method, electronic device, storage medium and program product provided by the present application relate to the field of autonomous driving technology. The method includes: obtaining the motion data and traffic scene data of the target vehicle in the target traffic scene, the motion data is determined based on the first motion trajectory data and / or posture of the target vehicle in the target traffic scene, and the traffic scene data is collected by the target vehicle; according to the traffic scene data, the scene motion information corresponding to each first traffic participant of the target vehicle is determined, and the scene motion information is determined based on the second motion trajectory data and / or posture of the corresponding first traffic participant in the target traffic scene; the first traffic participant includes objects other than the target vehicle in the target traffic scene; at least according to the motion data and the scene motion information corresponding to each associated traffic participant, the target traffic scene is constructed. The present application determines the scene motion information corresponding to each first traffic participant of the target vehicle according to the traffic scene data, fills the acquisition blind spot caused by the limitation of the acquisition perspective, and improves the scope of the constructed target traffic scene and the diversity of the detection target; according to the motion data and the scene motion information corresponding to each associated traffic participant, the target traffic scene is constructed to improve the coverage rate of the constructed target traffic scene. BRIEF DESCRIPTION OF THE DRAWINGS

[0093] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0094] Figure 1 Schematic diagram of the process of the traffic scene construction method provided in the embodiment of the present application Figure 1 ;

[0095] Figure 2 A schematic diagram of motion data provided by an embodiment of the present application;

[0096] Figure 3 A schematic diagram of first motion data provided by an embodiment of the present application;

[0097] Figure 4 A schematic diagram of a target traffic scenario provided in an embodiment of the present application;

[0098] Figure 5 A schematic diagram of the occupied space provided by an embodiment of the present application;

[0099] Figure 6 A schematic diagram of a traffic participant object provided in an embodiment of the present application;

[0100] Figure 7 A schematic diagram of scene motion information provided by an embodiment of the present application;

[0101] Figure 8 A schematic diagram of the path planning provided in the embodiment of the present application;

[0102] Fig. 9 A schematic diagram of the structure of the traffic scene construction device provided in this application;

[0103] Fig.10 A schematic diagram of the structure of the model determination device provided in this application;

[0104] Fig.11 A schematic diagram of the structure of the path planning device provided in this application;

[0105] Fig.12 A schematic diagram of the structure of an electronic device provided in one embodiment of the present application.

[0106] The above drawings have shown clear embodiments of the present application, which will be described in more detail later. These drawings and text descriptions are not intended to limit the scope of the present application in any way, but to illustrate the concept of the present application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0107] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0108] In response to the above problems, the present application provides a method for constructing a traffic scene, which determines the scene motion information corresponding to each first traffic participant of the target vehicle based on the traffic scene data, wherein the scene motion information is determined based on the second motion trajectory data and / or posture of the corresponding first traffic participant in the target traffic scene, thereby compensating for the blind spots caused by the limitation of the vehicle acquisition viewing angle, and improving the coverage of the target traffic scene and the diversity of the detection targets. Then, the target traffic scene is constructed by combining the motion data and the scene motion information corresponding to each associated traffic participant to improve the coverage of the target traffic scene.

[0109] The technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems are described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0110] Figure 1 Schematic diagram of the process of the traffic scene construction method provided in the embodiment of the present application Figure 1 ,like Figure 1 As shown, the method includes:

[0111] S101. Obtain motion data and traffic scene data of a target vehicle in a target traffic scene, wherein the motion data is determined based on first motion trajectory data and / or posture of the target vehicle in the target traffic scene, and the traffic scene data is collected by the target vehicle.

[0112] Among them, the target vehicle refers to a vehicle that can collect motion data and traffic scene data. The target vehicle can be any one or any combination of a private car, a taxi, a company's dedicated collection vehicle, and a crowdsourcing vehicle. The embodiment of the present application does not limit the type of the target vehicle.

[0113] Furthermore, the first motion trajectory data refers to the motion trajectory data of the target vehicle itself; and the traffic scene data refers to the data collected by the target vehicle for constructing the target traffic scene.

[0114] As an embodiment, those skilled in the art may set the target traffic scene according to actual needs, such as but not limited to setting the target traffic scene based on any information such as the travel time of the target vehicle, a specific traffic environment / scene, a geographical area, a specific building, a road area, etc. By providing different methods for determining the target traffic scene, the flexibility of the target traffic scene method is improved, and thus a variety of target traffic scenes can be constructed; for example, the target traffic scene can be determined based on at least one of the following scene determination methods A1 to A5:

[0115] Scene determination method A1: The traffic scene that the target vehicle passes through during the target time period is determined as the target traffic scene; for example, the traffic scene that the company's dedicated collection vehicle passes through between 10 a.m. and 10:30 a.m. is determined as the target traffic scene.

[0116] Scene determination method A2: determining a preset traffic scene as a target traffic scene; for example, determining a high-speed scene as a target traffic scene.

[0117] Scene determination method A3: determining a preset administrative district as a target traffic scene; for example, determining Y district in X city as a target traffic scene.

[0118] Scene determination method A4: determining the target traffic scene based on a specific building; for example, determining the range covered by a circle with Hospital X as the center and a preset distance as the radius as the target traffic scene.

[0119] Scene determination method A5: The target traffic scene is determined by a preset road; for example, a section of Provincial Highway Y located in Region X is determined as the target traffic scene.

[0120] As an embodiment, the traffic scene data obtained in step S101 includes at least one of images, videos, point clouds, and trajectories; that is, the form of expression of the traffic scene data collected by the target vehicle depends on the collection device of the target vehicle, wherein the collection device may be a camera, a lidar, etc., and the corresponding traffic scene data includes at least one of images, videos, point clouds, and trajectories.

[0121] Optionally, when the target traffic scene is a highway ramp, the target vehicle's own motion trajectory data and traffic scene data in the highway ramp can be obtained by the company's dedicated collection vehicle.

[0122] S102. Determine, based on the traffic scene data, scene motion information corresponding to each first traffic participant of the target vehicle, wherein the scene motion information is determined based on the second motion trajectory data and / or posture of the corresponding first traffic participant in the target traffic scene; the first traffic participant objects include objects in the target traffic scene except the target vehicle.

[0123] In this step, it can be understood that after determining the traffic scene data in S101, the scene motion information corresponding to each first traffic participant of the target vehicle needs to be determined according to the traffic scene data. The specific implementation method of determining the scene motion information corresponding to each first traffic participant of the target vehicle can be set according to actual needs.

[0124] In one implementation, an optical flow method is used on traffic scene data to determine scene motion information corresponding to each first traffic participant of the target vehicle, wherein the optical flow method includes any one or any combination of Lucas Kanade sparse optical flow algorithm, Gunnar Farneback dense optical flow algorithm, etc.

[0125] In another implementation, a scene flow method is used on the traffic scene data to determine the scene motion information corresponding to each first traffic participant of the target vehicle.

[0126] In another implementation, a deep learning method is used on the traffic scene data to determine the scene motion information corresponding to each first traffic participant of the target vehicle, wherein the deep learning method includes FlowNet, LiteFlowCOccAwareFlow, and the like.

[0127] It should be noted that the first traffic participant object is other objects that enter the target vehicle's collection field of view within the target traffic scene, etc. The other objects may include but are not limited to any one or any combination of vehicles, pedestrians, poles, traffic signs (such as but not limited to lane lines, traffic lights or zebra crossings, etc.), etc., and the first traffic participant object can be considered to include objects other than the target vehicle in the target traffic scene.

[0128] S103: construct a target traffic scene at least based on the motion data and the scene motion information corresponding to each associated traffic participant object.

[0129] By integrating the motion data and the scene motion information corresponding to each associated traffic participant, a target traffic scene with a high coverage rate is constructed. The specific implementation method of constructing the target traffic scene can be set according to actual needs.

[0130] In one implementation, motion data and scene motion information corresponding to each associated traffic participant object are input into scene construction software to complete the construction of the target traffic scene.

[0131] The embodiment of the present application determines the scene motion information corresponding to each first traffic participant of the target vehicle based on the traffic scene data, fills the acquisition blind spot caused by the limitation of the acquisition viewing angle, and improves the scope of the constructed target traffic scene and the diversity of the detection targets; constructs the target traffic scene according to the motion data and the scene motion information corresponding to each related traffic participant to improve the coverage of the constructed target traffic scene.

[0132] On the basis of the above-mentioned embodiment, according to the traffic scene data, the scene motion information corresponding to each first traffic participant of the target vehicle is determined, including: performing target recognition on the traffic scene data, and determining each recognized target as a first traffic participant; identifying the current motion information of each first traffic participant in the target traffic scene based on the traffic scene data; determining the current motion information of each first traffic participant in the target traffic scene as the scene motion information corresponding to each first traffic participant; or estimating the inferred motion information of each first traffic participant in the associated scene of the target traffic scene as the scene motion information corresponding to each first traffic participant based on the current motion information of each first traffic participant in the target traffic scene; or determining the inferred motion information of each first traffic participant in the associated scene of the target traffic scene and the current motion information of each first traffic participant in the target traffic scene as the scene motion information corresponding to each first traffic participant; wherein the associated scene is the historical scene before each first traffic participant enters the target traffic scene, and / or the associated scene is the future scene after each first traffic participant exits the target traffic scene.

[0133] In this embodiment, it can be understood that target recognition is performed on the traffic scene data, and the recognized target is determined as the first traffic participant object, wherein the specific implementation method of target recognition on the traffic scene data can be set according to actual needs, for example, target recognition is performed on the traffic scene data using a target recognition algorithm. Based on the traffic scene data, the current motion information of each first traffic participant object in the target traffic scene is identified.

[0134] Furthermore, the specific implementation method of determining the scene motion information corresponding to each first traffic participant of the target vehicle can be set according to actual needs.

[0135] In one implementation, the current motion information of each first traffic participant in the target traffic scene is determined as the scene motion information corresponding to each first traffic participant. In this implementation, the constructed target traffic scene includes the current motion information of each first traffic participant in the target traffic scene.

[0136] In another implementation, based on the current motion information of each first traffic participant in the target traffic scene, the inferred motion information of each first traffic participant in the associated scene of the target traffic scene is estimated as the scene motion information corresponding to each first traffic participant. Optionally, an improved Kalman filter method, and / or a vehicle motion model, and / or a Long Short-Term Memory (LSTM) network method integrating temporal information is used to estimate the inferred motion information of each first traffic participant in the associated scene of the target traffic scene, wherein the inferred motion information may be historical motion information and / or future motion information. In this implementation, the constructed target traffic scene includes the historical motion information and / or future motion information of each first traffic participant in the associated scene of the target traffic scene.

[0137] In another implementation, the inferred motion information of each first traffic participant in the associated scene of the target traffic scene and the current motion information of each first traffic participant in the target traffic scene are determined as the scene motion information corresponding to each first traffic participant. Among them, the method for determining the inferred motion information has been described in detail in the previous embodiment, so the embodiment of the present application will not be repeated here. In this implementation, the constructed target traffic scene includes the current motion information of each first traffic participant in the target traffic scene, the historical motion information and / or future motion information of each first traffic participant in the associated scene of the target traffic scene.

[0138] It should be noted that the associated scene is a historical scene before each first traffic participant enters the target traffic scene, and / or the associated scene is a future scene after each first traffic participant exits the target traffic scene.

[0139] The embodiments of the present application provide different methods for determining scene motion information corresponding to each first traffic participant of the target vehicle, thereby meeting different needs of users and improving the flexibility of the target traffic scene method.

[0140] Furthermore, in order to improve the accuracy of the target traffic scene construction method, it is necessary to correct and optimize the motion data and / or scene motion information. When determining whether it is necessary to correct and optimize the motion data and / or scene motion information, it can be set according to actual needs.

[0141] In one implementation, only the motion data is corrected.

[0142] In yet another implementation, only the scene motion information is corrected.

[0143] In another implementation, both the motion data and the scene motion information are corrected.

[0144] One point that needs to be explained is that the method used to correct motion data is similar to the method used to correct scene motion information. Therefore, the embodiment of the present application takes the correction of both motion data and scene motion information as an example to explain how to construct a target traffic scene based on the corrected motion data and scene motion information.

[0145] Furthermore, S103 describes constructing a target traffic scene at least based on motion data and scene motion information corresponding to each associated traffic participant object, including: correcting the motion data based on an electronic map to obtain first motion data; correcting the scene motion information based on an electronic map to obtain first scene motion information; and constructing a target traffic scene based on the first motion data and the first scene motion information.

[0146] In this embodiment, it can be understood that when determining the first motion data, it is necessary to first obtain the motion data of the target vehicle in the target traffic scene, for example, first obtain the motion data of the enterprise's dedicated collection vehicle in the time period from 10:00 to 10:30 and on the highway ramp, Figure 2 A schematic diagram of motion data provided by an embodiment of the present application, Figure 2 traj_list in represents motion data; CV_Pos_t0, CV_Pos_t1, and CV_Pos_t2 represent the positions of the target vehicle at t0, t1, and t2, respectively.

[0147] Based on the electronic map, the acquired motion data is corrected and optimized to obtain the first motion data of the target vehicle in the target traffic scene. The type of electronic map can be set according to actual needs. For example, a high-definition map (HDMap) is selected to correct and optimize the acquired motion data to obtain the first motion data. The first motion data can be found in Figure 3 , Figure 3 A schematic diagram of first motion data provided in an embodiment of the present application, Figure 3 traj_list_opt in represents the first motion data; CV_Pos_t0, CV_Pos_t1, and CV_Pos_t2 represent the positions of the target vehicle at t0, t1, and t2, respectively.

[0148] After completing the correction operation of the scene motion information and motion data, it is necessary to construct the target traffic scene based on the corrected scene motion information and the corrected motion data.

[0149] The embodiment of the present application corrects motion data and scene motion information based on an electronic map, and constructs a target traffic scene according to the corrected scene motion information and the corrected motion data, thereby improving the accuracy of constructing the target traffic scene.

[0150] In some embodiments, based on the electronic map, the motion data is corrected to obtain the first motion data, including: obtaining the initial position coordinates of the target vehicle based on GNSS, and obtaining the scene image captured by the target vehicle under the initial position coordinates; extracting the road elements in the scene image; determining the feature difference matrix based on the road elements in the scene image and the road elements in the electronic map; and correcting the motion data based on the feature difference matrix to obtain the first motion data.

[0151] Optionally, the initial position coordinates of the target vehicle under GNSS are obtained, and the scene image captured by the camera of the target vehicle is obtained under the initial position coordinates. The road elements in the scene image are extracted using an extraction algorithm, wherein the road elements include poles, lane lines, etc. The road elements in the scene image are matched with the road elements in the electronic map to determine a feature difference matrix. According to the determined feature difference matrix, the motion data is corrected and optimized to obtain the first motion data.

[0152] Furthermore, according to the determined characteristic difference matrix, the motion data is corrected and optimized, and the first motion data can be obtained by the following formula:

[0153] Traj_list_opt=Traj_list*RT

[0154] Among them, Traj_list_opt represents the first motion data; Traj_list represents the motion data; RT represents the feature difference matrix.

[0155] The embodiment of the present application corrects and optimizes the motion data according to a determined feature difference matrix to obtain first motion data, thereby improving the accuracy of constructing the target traffic scene.

[0156] Furthermore, in some embodiments, motion data is corrected according to a feature difference matrix to obtain first motion data, including: determining the position corresponding to any moment within a set time period, and performing the following processing on the motion data associated with the position: using the corrected first initial motion data as the initial value, and correcting the second initial motion data of the target vehicle according to the feature difference matrix, the first initial motion data being the motion data associated with the position corresponding to the current moment, and the second initial motion data being the motion data associated with the position corresponding to the next moment; and obtaining the first motion data based on the corrected motion data at all moments.

[0157] Optionally, assuming that the current moment is moment t-1 and the next moment is moment t, the correction of the first motion trajectory data associated with the position corresponding to moment t is achieved by the following formula:

[0158] Pt '=F(f(hdmap),λP t-1 ',ωP t )

[0159] Where Pt' represents the corrected motion data associated with the position corresponding to time t; f(hdmap) represents the feature difference matrix; λ and ω represent the influencing factors; P t Represents the motion data associated with the position corresponding to time t; Pt- 1 ' represents the corrected motion data associated with the position corresponding to time t-1.

[0160] In the embodiment of the present application, each correction of motion data is based on the optimized motion data of the previous moment as the initial value, which can reduce error accumulation and thus improve the accuracy of constructing the target traffic scene.

[0161] Furthermore, in some embodiments, the traffic scene construction method provided in the embodiments of the present application also includes: determining the target lane of the target vehicle in the electronic map according to the initial position coordinates of the target vehicle based on GNSS; and correcting the motion data according to the target lane.

[0162] In this embodiment, it can be understood that the motion data can also be corrected by determining the target lane of the target vehicle in the electronic map, wherein the target lane of the target vehicle in the electronic map is determined by comparing the initial position coordinates of the target vehicle under the global navigation satellite system with the position information of the lane in the electronic map.

[0163] After the target lane is determined, the motion data of the target vehicle is adjusted to the precise lane centerline using map matching technology and / or filtering algorithms to correct the deviation.

[0164] It should be noted that the correction of the motion data based on the target lane can be performed before the correction operation on the motion data based on the electronic map, or it can be performed after the correction operation on the motion data based on the electronic map, wherein the specific order can be selected according to actual needs.

[0165] The embodiment of the present application uses the target lane to correct the actual position of the target vehicle, thereby reducing the deviation caused by environmental interference, etc., so that more accurate first motion data can be obtained.

[0166] Furthermore, the principle of correcting scene motion information is similar to the principle of correcting motion data, so the specific principle of correcting scene motion information is not further described in the embodiments of the present application.

[0167] Furthermore, in some embodiments, a target traffic scene is constructed based on at least the motion data and the scene motion information corresponding to each associated traffic participant, including: obtaining the occupied space of the target vehicle in the target traffic scene; constructing the target traffic scene based on the motion data, the scene motion information and the occupied space, wherein a schematic diagram of the constructed target traffic scene can be seen in Figure 4 , Figure 4 A schematic diagram of a target traffic scenario provided in an embodiment of the present application, Figure 4 Where a represents the motion trajectory data of the target vehicle, b represents the motion trajectory data of the traffic participant object, and c represents the motion trajectory data of another traffic participant object.

[0168] In this embodiment, the occupied space can be understood as the space where the target vehicle appears in the target traffic scene, and the occupied space can be the area where the target vehicle appears in the entire target traffic scene; for example, when the traffic scene where vehicle A (target vehicle) passes through from 8:00 to 8:10 for 10 minutes is determined as the target traffic scene, the occupied space of vehicle A in the target traffic scene can be the sum of the areas where vehicle A appears in the target traffic scene within the 10 minutes. Figure 5 A schematic diagram of the occupied space provided in the embodiment of the present application, Figure 5 The different boxes in the figure represent the projection area of ​​the target vehicle on the ground at different sampling times, θ i Represents the inclination angle between the target vehicle and the ground at different sampling times.

[0169] For example, in this example, the basic size information of vehicle A, such as the length and width of the vehicle, can be obtained first. Next, the precise position and posture of vehicle A in the target traffic scene are obtained (for example, the precise position and posture of the vehicle are obtained using GPS and an inertial measurement unit); then, the projection area of ​​vehicle A on the ground is calculated based on the size of vehicle A (such as the length and width of vehicle A) and posture. Specifically, assuming that the length of vehicle A is L and the width of vehicle A is W, the posture is described by the angle θ between vehicle A and the ground. When the inclination angle θ between vehicle A and the ground is equal to 0°, the projection area of ​​vehicle A on the ground is the product of the length and width of vehicle A. However, in actual traffic scenes, the posture of vehicle A will undergo various changes. For example, when turning or avoiding other vehicles, the inclination angle θ between vehicle A and the ground is not equal to 0°. At this time, the calculation of the projection area of ​​vehicle A on the ground needs to be performed using trigonometric functions. The effective projection length of the length of vehicle A when perpendicular to the ground becomes L×|cosθ|, the effective projection length of the width of vehicle A when perpendicular to the ground becomes W×|sinθ|, and the projection area of ​​vehicle A on the ground is L×|cosθ|×W×|sinθ|.

[0170] Furthermore, since the position and posture of vehicle A are in a constantly changing state during driving, in order to more accurately describe the space occupied by vehicle A during driving, it is necessary to set a sampling time interval to sample and record the position and posture of vehicle A. For each sampling moment, the projection area of ​​vehicle A at this time is calculated according to the above method. For example, when vehicle A performs an evasive operation during driving, the inclination angle θ between vehicle A and the ground will change, and a new θ is needed to recalculate the projection area. Based on different sampling moments, the projection areas that appear in the target traffic scene are accumulated to obtain the occupied space of vehicle A.

[0171] In summary, the occupied space can be understood as the sum of the projected areas of the target vehicles that will appear in the target traffic scene at different sampling times.

[0172] Furthermore, the occupied space includes the posture of the target vehicle at each sampling moment and the sum of all projected areas of the target vehicle in the target traffic scene.

[0173] After determining the space occupied by the target vehicle within a set time period, the motion data, scene motion information, and occupied space are input into the scene construction software to complete the construction of the target traffic scene.

[0174] The embodiment of the present application can describe in detail the dynamic changes of the traffic environment in the target traffic scene by analyzing motion data, scene motion information and occupied space. This integration helps to simulate real traffic conditions and optimize traffic management, and also helps to provide environmental perception and decision support for the autonomous driving system, thereby improving the overall efficiency and safety of the traffic system.

[0175] Furthermore, with the rapid development of autonomous driving technology, achieving accurate, intelligent and safe path planning has become a core requirement. At present, the exhaustive method is often used to train the path planning ability of the vehicle, that is, to collect various scene data as much as possible, so that the vehicle can be exposed to diverse traffic conditions in simulation or actual testing and optimize decision-making. However, due to the exponential growth of the complexity of traffic scenes, it is impossible to exhaustively enumerate all traffic scenes, which leads to problems in path planning when the vehicle encounters unlearned traffic scenes during actual driving.

[0176] In order to solve the above problems, an embodiment of the present application also provides a model determination method, which includes: obtaining scene data corresponding to the target traffic scene, the target traffic scene is constructed by the traffic scene construction method described in the previous embodiment; using the scene data to train or optimize the self-vehicle prediction planning model, and the self-vehicle prediction planning model is used for vehicle path planning.

[0177] This embodiment describes a systematic process for acquiring scene data from a target traffic scene and using the scene data to train or optimize an ego vehicle prediction planning model to improve the vehicle's path planning.

[0178] Specifically, first, the traffic scene construction method described in the previous embodiment generates a target traffic scene. This scene is based on the collected multiple data sources, including vehicle motion trajectory data, lane lines, etc.

[0179] After the target traffic scene is constructed, the scene data corresponding to the target traffic scene needs to be obtained. These data describe the dynamic and static elements in the scene in detail, such as lane usage and vehicle flow in different time periods, etc.

[0180] The above scenario data is used to train or optimize the self-driving vehicle prediction planning model. The self-driving vehicle prediction planning model provides accurate path planning and decision support for the vehicle. By training on real and complex scenario data, the self-driving vehicle prediction planning model can learn to identify and adapt to changing traffic patterns, improving the prediction accuracy and robustness of the self-driving vehicle prediction planning model in practical applications.

[0181] Furthermore, the trained or optimized self-driving prediction planning model can help vehicles plan routes efficiently in complex traffic environments. It can not only select the best route to reduce driving time, but also adjust driving strategies in advance by predicting potential traffic problems to ensure the safety of drivers.

[0182] The embodiments of the present application utilize these scenario data to train or optimize the vehicle prediction planning model to enhance the vehicle's intelligent path planning capabilities.

[0183] Based on the above embodiments, an embodiment of the present application provides a path planning method, which is applied to a vehicle. The path planning method includes: acquiring scene data; inputting the scene data into a self-vehicle prediction planning model to obtain a path planning for the vehicle. The self-vehicle prediction planning model is determined by the model determination method described in the above embodiments.

[0184] This embodiment describes the acquisition of scene data and the use of these scene data for path planning. Specifically, firstly, scene data is acquired, which includes road information, traffic flow, environmental conditions, etc., which fully reflect the dynamic and static elements of the current traffic scene. Then, these traffic scene data are input into the self-driving prediction planning model, which is determined by the method described in the above embodiment. The self-driving prediction planning model is capable of processing complex scene data, analyzing the current traffic status and predicting future traffic changes. Through the above analysis, the self-driving prediction planning model can identify a better path.

[0185] Furthermore, the self-driving prediction planning model can generate a vehicle path plan based on the input scene data. The path plan provides detailed navigation guidance to help the vehicle drive in a more optimal way, reduce the vehicle's driving time, and improve the safety of the vehicle.

[0186] In summary, the embodiments of the present application provide intelligent path planning solutions for vehicles by utilizing scene data and the self-vehicle prediction planning model, thereby improving the efficiency and reliability of the transportation system.

[0187] Next, taking vehicle A as the target vehicle as an example, the traffic scene construction method provided by the embodiment of the present application is explained, and the method includes the following steps:

[0188] 1. Obtaining the motion data and traffic scene data of vehicle A on the highway corresponding to the first time period, wherein the first time period is from 10:00 to 10:30, and the motion data includes the motion trajectory data and / or posture of vehicle A itself;

[0189] 2. Based on the high-definition map, the motion data is corrected and optimized to obtain the first motion data;

[0190] 3. Perform target recognition and trajectory monitoring on the traffic scene data in step 1 to determine the scene motion information of the traffic participants in the field of view of the video stream of vehicle A and the first time period. For details, see Figure 6 , Figure 7 , Figure 6 A schematic diagram of a traffic participant object provided in an embodiment of the present application, Figure 7 A schematic diagram of scene motion information provided by an embodiment of the present application, wherein: Figure 6 Where a1 represents the position of vehicle A at time t-1, b1 represents the position of the traffic participant at time t, c1 represents the position of the traffic participant in the field of view of the video stream of vehicle A at time t-1, and d1 represents the position of the traffic participant in the field of view of the video stream of vehicle A at time t. The participants include vehicles, pedestrians, other movable objects, etc. other than vehicle A, that is, other vehicles, pedestrians, other movable objects, etc. that enter the acquisition field of view of vehicle A; scene motion information includes motion parameters and motion trajectory data, and motion parameters include position, speed, acceleration, etc.;

[0191] 4. Based on the scene motion information of the determined traffic participants, for the traffic participants within the first time period or outside the field of view but within the highway scene segment, the scene motion information of the corresponding traffic participants is predicted by using methods such as improved Kalman filtering, vehicle motion model-based or LSTM network integrating time series information, including historical scene motion information and / or future scene motion information; in this way, the spatiotemporal posture information of vehicle A and other traffic participants in the entire highway traffic scene segment corresponding to the first time period completely covers the entire time period and the entire road segment;

[0192] 5. Based on the high-definition map, correct and optimize the first motion data and the scene motion information of the traffic participants in the first time period in the entire highway traffic scene segment;

[0193] 6. Obtain the occupied space of vehicle A in the first time period;

[0194] 7. Construct a traffic scene on the highway based on the occupied space, the corrected first motion data, and the corrected scene motion information of the traffic participants in the first time period in the entire highway traffic scene segment.

[0195] Further, the scene data corresponding to the constructed traffic scene is obtained, and the scene data is used to train or optimize the self-vehicle prediction planning model. Further, the scene data is obtained, and the scene data is input into the self-vehicle prediction planning model to obtain the path planning of vehicle A. For a specific path planning diagram, see Figure 8 , Figure 8 A schematic diagram of the path planning provided for an embodiment of the present application.

[0196] In summary, the embodiment of the present application proposes a method for constructing a scene library based on high-definition maps and crowdsourced trajectory data. By extracting the trajectories or occupancy grids of other dynamic or static obstacles such as other vehicles and pedestrians in the vehicle based on vehicles at different times, and superimposing the high-definition map, the motion trajectory data of all traffic participants in the scene within the time series can be provided, and the corresponding traffic scene can be constructed based on the motion trajectory data.

[0197] Furthermore, the scene data corresponding to the traffic scene is imported into the self-vehicle prediction planning model, and the relationship between the vehicle and other vehicles in the scene is trained through random or customized autonomous driving vehicle simulation to predict the effect of the planning decision algorithm. Furthermore, the method provided in the embodiment of the present application reduces the extraction of scenes through field vehicle road testing, and can also greatly improve the coverage of scenes and the dynamic monitoring of various special scenes at multiple times and changes.

[0198] The following is an embodiment of the device of the present application, which can be used to execute the embodiment of the method of the present application. For details not disclosed in the embodiment of the device of the present application, please refer to the embodiment of the method of the present application.

[0199] Fig. 9 A schematic diagram of the structure of the traffic scene construction device provided in this application, such as Fig. 9 As shown, the traffic scene construction device 900 provided in this embodiment includes:

[0200] An acquisition module 901 is used to acquire motion data and traffic scene data of a target vehicle in a target traffic scene, wherein the motion data is determined based on first motion track data and / or posture of the target vehicle in the target traffic scene, and the traffic scene data is collected by the target vehicle;

[0201] A determination module 902 is used to determine, based on the traffic scene data, scene motion information corresponding to each first traffic participant of the target vehicle, wherein the scene motion information is determined based on the second motion trajectory data and / or posture of the corresponding first traffic participant in the target traffic scene; the first traffic participant objects include objects other than the target vehicle in the target traffic scene;

[0202] The construction module 903 is used to construct a target traffic scene based on at least the motion data and the scene motion information corresponding to each associated traffic participant object.

[0203] In a possible implementation, the determination module 902 is specifically configured to:

[0204] Determine the traffic scene that the target vehicle passes through during the target period as the target traffic scene;

[0205] and / or, determining a preset traffic scene as a target traffic scene;

[0206] And / or, the traffic scene data includes at least one of an image, a video, a point cloud, and a trajectory.

[0207] In a possible implementation, the determination module 902 is specifically configured to:

[0208] Performing target recognition on the traffic scene data, and determining each recognized target as a first traffic participant;

[0209] Identifying current motion information of each first traffic participant in the target traffic scene based on the traffic scene data;

[0210] Determine the current motion information of each first traffic participant in the target traffic scene as the scene motion information corresponding to each first traffic participant; or

[0211] Based on the current motion information of each first traffic participant in the target traffic scene, the inferred motion information of each first traffic participant in the associated scene of the target traffic scene is estimated as the scene motion information corresponding to each first traffic participant; or

[0212] The inferred motion information of each first traffic participant in the associated scene of the target traffic scene and the current motion information of each first traffic participant in the target traffic scene are determined as the scene motion information corresponding to each first traffic participant; wherein

[0213] The associated scene is a historical scene before each first traffic participant enters the target traffic scene, and / or the associated scene is a future scene after each first traffic participant exits the target traffic scene.

[0214] In one possible implementation, the construction module 903 is specifically used to: correct the motion data based on the electronic map to obtain the first motion data; correct the scene motion information based on the electronic map to obtain the first scene motion information; and construct the target traffic scene according to the first motion data and the first scene motion information.

[0215] In a possible implementation, the traffic scene construction device also includes a processing module (not shown), which is specifically used to: obtain the initial position coordinates of the target vehicle based on GNSS, and obtain the scene image collected by the target vehicle under the initial position coordinates; extract road elements in the scene image; determine a feature difference matrix based on the road elements in the scene image and the road elements in the electronic map; and correct the motion data based on the feature difference matrix to obtain first motion data.

[0216] In a possible implementation manner, the processing module is specifically used for:

[0217] Determine the position corresponding to any time within the set time period, and perform the following processing on the motion data associated with the position: take the corrected first initial motion data as the initial value, and correct the second initial motion data of the target vehicle according to the feature difference matrix, the first initial motion data is the motion data associated with the position corresponding to the current time, and the second initial motion data is the motion data associated with the position corresponding to the next time;

[0218] The first motion data is obtained according to the motion data corrected at all times.

[0219] In a possible implementation, the processing module is specifically used to: determine a target lane of the target vehicle in the electronic map according to the initial position coordinates of the target vehicle based on GNSS; and correct the motion data according to the target lane.

[0220] In a possible implementation, the construction module 903 is further used to: obtain the occupied space of the target vehicle in the target traffic scene; and construct the target traffic scene according to the motion data, the scene motion information, and the occupied space.

[0221] Fig.10 A schematic diagram of the structure of the model determination device provided in this application, such as Fig.10 As shown, the model determination device 1000 provided in this embodiment includes:

[0222] An acquisition module 1001 is used to acquire scene data corresponding to a target traffic scene, where the target traffic scene is constructed through the above-mentioned various possible implementation methods;

[0223] The processing module 1002 is used to use the scene data to train or optimize the ego vehicle prediction planning model, and the ego vehicle prediction planning model is used for vehicle path planning.

[0224] Fig.11 A schematic diagram of the structure of the path planning device provided in this application, such as Fig.11 As shown, the path planning device 1100 provided in this embodiment includes:

[0225] An acquisition module 1101 is used to acquire scene data;

[0226] The planning module 1102 is used to input the scene data into the ego vehicle prediction planning model to obtain the path planning of the vehicle. The ego vehicle prediction planning model is determined by the various possible implementation methods described above.

[0227] The traffic scene construction device, model determination device, and path planning device provided in this embodiment can execute the method provided in the above method embodiment. The implementation principles and technical effects are similar, and are not described in detail in this embodiment.

[0228] It should be noted that it should be understood that the division of the various modules of the above device is only a division of logical functions. In actual implementation, they can be fully or partially integrated into one physical entity, or they can be physically separated. And these modules can all be implemented in the form of software called by processing elements; they can also be all implemented in the form of hardware; some modules can also be implemented in the form of software called by processing elements, and some modules can be implemented in the form of hardware. For example, the processing module can be a separately established processing element, or it can be integrated in a chip of the above device. In addition, it can also be stored in the memory of the above device in the form of program code, and called and executed by a processing element of the above device. The implementation of other modules is similar. In addition, these modules can be fully or partially integrated together, or they can be implemented independently. The processing element here can be an integrated circuit with signal processing capabilities. In the implementation process, each step of the above method or each module above can be completed by an integrated logic circuit of hardware in the processor element or instructions in the form of software.

[0229] For example, the above modules may be one or more integrated circuits configured to implement the above methods, such as one or more application specific integrated circuits (ASIC), or one or more microprocessors (DSP), or one or more field programmable gate arrays (FPGA). For another example, when a module is implemented in the form of a processing element to allocate program code, the processing element may be a general-purpose processor, such as a central processing unit (CPU) or other processors that can call program code. For another example, these modules may be integrated together and implemented in the form of a system-on-a-chip (SOC).

[0230] Fig.12 This is a schematic diagram of the structure of an electronic device provided by an embodiment of the present application. Fig.12 As shown, the electronic device 1200 provided in the embodiment of the present application may include: a processor 1201, and a memory 1202 connected to the processor in communication, wherein:

[0231] Memory stores computer-executable instructions;

[0232] The processor executes the computer-executable instructions stored in the memory to implement the method described in the foregoing method embodiment.

[0233] It should be understood that the processor 1201 can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), etc. The general-purpose processor can be a microprocessor or the processor can be any conventional processor, etc. The steps of the method disclosed in the application can be directly embodied as being executed by a hardware processor, or executed by a combination of hardware and software modules in the processor. The memory 1202 may include a high-speed random access memory (RAM), and may also include non-volatile storage NVM (non-volatile memory), such as at least one disk storage, and can also be a USB flash drive, a mobile hard disk, a read-only memory, a disk or an optical disk, etc.

[0234] Optionally, the electronic device 1200 may further include a communication interface 1203. In a specific implementation, if the communication interface 1203, the memory 1202 and the processor 1201 are implemented independently, the communication interface 1203, the memory 1202 and the processor 1201 may be interconnected through a bus and communicate with each other. The bus may be an Industry Standard Architecture (ISA) bus, a Peripheral Component (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc., but it does not mean that there is only one bus or one type of bus.

[0235] Optionally, in a specific implementation, if the communication interface 1203, the memory 1202 and the processor 1201 are integrated on a chip, the communication interface 1203, the memory 1202 and the processor 1201 can communicate through an internal interface.

[0236] An embodiment of the present application also provides a computer-readable storage medium, in which computer-executable instructions are stored. When the computer-executable instructions are executed, they are used to implement the method described in any of the aforementioned embodiments.

[0237] It is understood that the computer readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read only memory (EEPROM), erasable programmable read only memory (EPROM), programmable read only memory (PROM), read only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk. The readable storage medium can be any available medium that can be accessed by a general or special computer.

[0238] An exemplary computer-readable storage medium is coupled to a processor so that the processor can read information from the computer-readable storage medium and write information to the computer-readable storage medium. Of course, the computer-readable storage medium can also be an integral part of the processor. The processor and the computer-readable storage medium can be located in an ASIC. Of course, the processor and the computer-readable storage medium can also exist in an electronic device as discrete components.

[0239] The above-mentioned integrated module implemented in the form of a software function module can be stored in a computer-readable storage medium. The above-mentioned software function module is stored in a computer-readable storage medium, including a number of instructions for enabling an electronic device (which can be a personal computer, a server, or a network device, etc.) or a processor to perform some steps of the method described in each embodiment of the present application.

[0240] An embodiment of the present application also provides a computer program product, including a computer program, which implements the method described in any of the above embodiments when executed.

[0241] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily required by the present application.

[0242] It should be further noted that, although the various steps in the flowchart are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless there is a clear description in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowchart may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these sub-steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of the sub-steps or stages of other steps.

[0243] In the above embodiments, the description of each embodiment has its own emphasis. For the part not described in detail in a certain embodiment, please refer to the relevant description of other embodiments. The technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, all possible combinations of the technical features in the above embodiments are not described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0244] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention disclosed herein. The present application is intended to cover any modification, use or adaptation of the present application, which follows the general principles of the present application and includes common knowledge or customary techniques in the art that are not disclosed in the present application. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present application are indicated by the following claims.

[0245] It should be understood that the present application is not limited to the precise structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. A traffic scene construction method, characterized in that: include: Acquire motion data and traffic scene data of a target vehicle in a target traffic scene, wherein the motion data is determined based on first motion track data and / or posture of the target vehicle in the target traffic scene, and the traffic scene data is collected by the target vehicle; Determine, according to the traffic scene data, scene motion information corresponding to each first traffic participant object of the target vehicle, wherein the scene motion information is determined based on second motion trajectory data and / or posture of the corresponding first traffic participant in the target traffic scene; the first traffic participant object includes objects in the target traffic scene other than the target vehicle; The target traffic scene is constructed at least based on the motion data and the scene motion information corresponding to each associated traffic participant object.

2. The method according to claim 1, characterized in that The method further comprises: Determining the traffic scene that the target vehicle passes through during the target period as the target traffic scene; and / or, determining a preset traffic scene as the target traffic scene; and / or, The traffic scene data includes at least one of an image, a video, a point cloud, and a trajectory.

3. The method according to claim 1, characterized in that The step of determining scene motion information corresponding to each first traffic participant of the target vehicle according to the traffic scene data includes: Performing target recognition on the traffic scene data, and determining each recognized target as the first traffic participant; Identifying current movement information of each first traffic participant in the target traffic scene based on the traffic scene data; Determining the current motion information of each first traffic participant object in the target traffic scene as the scene motion information corresponding to each first traffic participant object; or Based on the current motion information of each first traffic participant in the target traffic scene, the inferred motion information of each first traffic participant in the associated scene of the target traffic scene is estimated as the scene motion information corresponding to each first traffic participant; or The inferred motion information of each first traffic participant in the associated scene of the target traffic scene and the current motion information of each first traffic participant in the target traffic scene are determined as the scene motion information corresponding to each first traffic participant; wherein The associated scene is a historical scene before each first traffic participant enters the target traffic scene, and / or the associated scene is a future scene after each first traffic participant exits the target traffic scene.

4. The method according to any one of claims 1 to 3, characterized in that The step of constructing the target traffic scene at least according to the motion data and the scene motion information corresponding to each associated traffic participant object includes: Based on the electronic map, correcting the motion data to obtain first motion data; Based on the electronic map, correcting the scene motion information to obtain first scene motion information; The target traffic scene is constructed according to the first motion data and the first scene motion information.

5. The method according to claim 4, characterized in that The correcting the motion data based on the electronic map to obtain first motion data includes: Acquire the initial position coordinates of the target vehicle based on the global navigation satellite system GNSS, and acquire the scene image collected by the target vehicle under the initial position coordinates; Extracting road elements in the scene image; Determining a feature difference matrix based on the road elements in the scene image and the road elements in the electronic map; The motion data is corrected according to the feature difference matrix to obtain the first motion data.

6. The method according to claim 5, characterized in that The step of correcting the motion data according to the feature difference matrix to obtain the first motion data includes: Determine the position corresponding to any time within the set time period, and perform the following processing on the motion data associated with the position: take the corrected first initial motion data as the initial value, and correct the second initial motion data of the target vehicle according to the characteristic difference matrix, the first initial motion data is the motion data associated with the position corresponding to the current time, and the second initial motion data is the motion data associated with the position corresponding to the next time; The first motion data is obtained according to the motion data corrected at all times.

7. The method according to claim 4, characterized in that Also includes: Determining a target lane of the target vehicle in the electronic map according to the initial position coordinates of the target vehicle based on GNSS; The motion data is corrected according to the target lane.

8. The method according to any one of claims 1 to 3, characterized in that The step of constructing the target traffic scene at least according to the motion data and the scene motion information corresponding to each associated traffic participant object includes: Obtaining an occupied space of the target vehicle within the target traffic scene; The target traffic scene is constructed according to the motion data, the scene motion information and the occupied space.

9. A model determination method, characterized in that: include: Acquire scene data corresponding to a target traffic scene, wherein the target traffic scene is constructed by the traffic scene construction method according to any one of claims 1 to 8; The scenario data is used to train or optimize a self-vehicle prediction planning model, and the self-vehicle prediction planning model is used for path planning of the vehicle.

10. A path planning method, characterized in that: Applied to a vehicle, the path planning method includes: Get scene data; The scene data is input into a self-vehicle prediction planning model to obtain a path planning for the vehicle, wherein the self-vehicle prediction planning model is determined by the method described in claim 9.

11. An electronic device, characterized in that: include: Memory, processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory, so that the processor performs the method according to any one of claims 1-8, 9 or 10.

12. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are executed, they are used to implement the method according to any one of claims 1-8 or 9 or 10.

13. A computer program product, characterized in that The invention comprises a computer program, which implements the method according to any one of claims 1 to 8 or 9 or 10 when being executed.

Images