Methods, devices, electronic equipment and storage media for predicting routes of traffic participants

Abstract

This invention relates to the field of intelligent driving technology, specifically to a method, device, electronic device, and storage medium for predicting the path of traffic participants. It includes: acquiring target feature information corresponding to a target trailer and other feature information corresponding to other traffic participants within a preset range; generating a structural graph of the target trailer and other traffic participants based on the relationship between the target feature information and other feature information corresponding to other traffic participants; inputting the structural graph into a preset target graph neural network for training, and outputting the predicted path of the target trailer and other predicted paths corresponding to other traffic participants. This method fully considers the shape of the target trailer and other traffic participants, avoiding misjudgments due to incomplete information, which could lead to collisions or cautious stopping.

Description

Technical Field

[0001] This invention relates to the field of intelligent driving technology, specifically to a method, device, electronic device, and storage medium for predicting the path of traffic participants. Background Technology

[0002] During operation, autonomous vehicles need to interact with surrounding traffic participants and the road at all times. While most road information is static (except for traffic lights), traffic participants are constantly changing. Traditional deep learning and prediction algorithms can no longer meet the needs of dynamic trajectory prediction.

[0003] In existing technologies, graph neural networks are typically used for dynamic trajectory prediction. Each traffic participant is a node in the graph, and the information interaction relationships between traffic participants can be represented by graph edges. The driving scenario at each moment can be mapped to a graph. By utilizing the relationships between traffic participants in the formed graph, dynamic trajectory prediction is performed on the traffic participants.

[0004] However, current methods for predicting the trajectory of moving obstacles using graph neural networks treat traffic participants as point masses. The predicted position information only includes the x, y, and z coordinates relative to the map, ignoring the shape information of traffic participants. For vehicles that rely on the prediction results for motion planning, this incomplete information can lead to misjudgments, resulting in collisions or cautious stops. Summary of the Invention

[0005] In view of this, embodiments of the present invention provide a method for predicting the path of traffic participants, which aims to solve the problem that incomplete information may lead to misjudgments, resulting in collisions or cautious stopping.

[0006] According to a first aspect, embodiments of the present invention provide a method for predicting the routes of traffic participants, comprising:

[0007] Obtain target feature information corresponding to the target trailer within a preset range and other feature information corresponding to other traffic participants; target feature information is used to characterize the feature state of the vehicle body target box and the vehicle front target box corresponding to the target trailer, and other feature information is used to characterize the feature state of other boxes corresponding to other traffic participants.

[0008] Based on the relationship between the target feature information corresponding to the target trailer and other feature information corresponding to other traffic participants, a structural diagram of the target trailer and other traffic participants is generated.

[0009] The structure graph is input into a preset target graph neural network to train the structure graph and output the target predicted path corresponding to the target trailer and other predicted paths corresponding to other traffic participants.

[0010] The traffic participant path prediction method provided in this invention obtains target feature information corresponding to a target trailer and other feature information corresponding to other traffic participants within a preset range. The target feature information characterizes the feature states of the vehicle body target box and the vehicle front target box corresponding to the target trailer, while the other feature information characterizes the feature states of other boxes corresponding to other traffic participants. This demonstrates that the obtained target feature information corresponding to the target trailer can characterize the shape of the target trailer's front and body, and the other feature information corresponding to other traffic participants can also characterize the shape of other traffic participants. This ensures the accuracy of the obtained target feature information and other feature information. Based on the relationship between the target feature information corresponding to the target trailer and other feature information corresponding to other traffic participants, a structure graph corresponding to the target trailer and other traffic participants is generated, ensuring the accuracy of the generated structure graph. Then, the structure graph is input into a preset target graph neural network for training, outputting the target predicted path corresponding to the target trailer and other predicted paths corresponding to other traffic participants, ensuring the accuracy of the predicted target predicted path corresponding to the target trailer and other predicted paths corresponding to other traffic participants. The method described above fully considers the shapes of the target trailer and other traffic participants, without treating them as point masses. Therefore, it ensures that the predicted path for the target trailer includes not only its x, y, z coordinates but also its shape. Similarly, the predicted paths for other traffic participants include not only their x, y, z coordinates but also their shape. This avoids misjudgments due to incomplete information, which could lead to collisions or forced stops.

[0011] In conjunction with the first aspect, in the first embodiment of the first aspect, a structure graph is input into a preset target graph neural network to train the structure graph, and outputs the target predicted path corresponding to the target trailer and other predicted paths corresponding to other traffic participants, including:

[0012] Based on the relationships between the target trailer and other traffic participants in the structural diagram, determine the driving priority between the target trailer and other traffic participants.

[0013] The structure graph and driving priority are input into a preset target graph neural network. The structure graph and driving priority are trained to output the target predicted path corresponding to the target trailer and other predicted paths corresponding to other traffic participants.

[0014] The traffic participant path prediction method provided in this invention determines the driving priority between each pair of target trailers and other traffic participants based on the pairwise relationships between them in a structure graph, ensuring the accuracy of the determined driving priority. The structure graph and driving priority are input into a preset target graph neural network for training, outputting the target predicted path for the target trailer and other predicted paths for other traffic participants. This method fully considers the driving priority between each pair of target trailers and other traffic participants, ensuring the accuracy of the output target predicted path and other predicted paths. It avoids misjudgments due to incomplete information, which could lead to collisions or cautious stopping.

[0015] In conjunction with the first embodiment of the first aspect, in the second embodiment of the first aspect, determining the driving priority between each pair of the target trailer and other traffic participants based on the pairwise relationships between the target trailer and other traffic participants in the structural diagram includes:

[0016] Obtain the driving speed, obstacle type, and current state of the target trailer and other traffic participants; where the current state is used to represent the angle between the front of the target trailer and the body of the other traffic participants.

[0017] Based on the relative relationships between the target trailer and other traffic participants’ respective driving speeds, obstacle types, and current states, the driving priority between each pair of the target trailer and other traffic participants is determined.

[0018] The traffic participant path prediction method provided in this invention obtains the driving speed, obstacle type, and current state of the target trailer and other traffic participants. Then, based on the relative relationships between the driving speed, obstacle type, and current state of the target trailer and other traffic participants, the driving priority between each pair of the target trailer and other traffic participants is determined, ensuring the accuracy of the determined driving priority between each pair of the target trailer and other traffic participants.

[0019] In conjunction with the second embodiment of the first aspect, in the third embodiment of the first aspect, the driving priority between each pair of the target trailer and other traffic participants is determined based on the relative relationships between their respective driving speeds, obstacle types, and current states, including:

[0020] For any two traffic participants among the target trailer and other traffic participants, calculate the speed priority based on the respective driving speeds of the two traffic participants;

[0021] Calculate obstacle category priority based on the obstacle category corresponding to each of the two traffic participants;

[0022] Calculate the priority of the current state based on the current state of each of the two traffic participants;

[0023] The driving priority between the target trailer and other traffic participants is determined based on the speed priority, obstacle category priority, and current state priority of any two traffic participants.

[0024] The traffic participant path prediction method provided in this invention calculates speed priority for any two traffic participants (target trailer and other traffic participants) based on their respective driving speeds, ensuring the accuracy of the calculated speed priority. It then calculates obstacle category priority based on the obstacle categories of the two traffic participants, ensuring the accuracy of the calculated obstacle category priority. Next, it calculates current state priority based on the current states of the two traffic participants, ensuring the accuracy of the calculated current state priority. Finally, based on the speed priority, obstacle category priority, and current state priority of any two traffic participants, the method determines the travel priority between each pair of target trailers and other traffic participants, ensuring the accuracy of the determined travel priority between each pair of target trailers and other traffic participants.

[0025] In conjunction with the first embodiment of the first aspect, in the fourth embodiment of the first aspect, the structure graph and driving priority are input into a preset target graph neural network, the structure graph and driving priority are trained, and the target predicted path corresponding to the target trailer and other predicted paths corresponding to other traffic participants are output, including;

[0026] The structure diagram and driving priority are input into a preset target graph neural network, which learns and encodes the structure diagram and driving priority to obtain encoded data.

[0027] Graph convolution is performed on the encoded data to generate feature vectors corresponding to the target trailer and other traffic participants within a preset time period in the future;

[0028] Analyze each feature vector and output the target predicted path for the target trailer and other predicted paths for other traffic participants.

[0029] The traffic participant path prediction method provided in this invention inputs a structure graph and driving priorities into a preset target graph neural network. The preset target graph neural network learns and encodes the structure graph and driving priorities to obtain encoded data, ensuring the accuracy of the obtained encoded data. Graph convolution is performed on the encoded data to generate feature vectors corresponding to the target trailer and other traffic participants within a preset future time period, ensuring the accuracy of the generated feature vectors. Each feature vector is analyzed to output the target predicted path corresponding to the target trailer and other predicted paths corresponding to other traffic participants, ensuring the accuracy of the output target predicted path corresponding to the target trailer and other predicted paths corresponding to other traffic participants.

[0030] In conjunction with the fourth embodiment of the first aspect, in the fifth embodiment of the first aspect, the feature vector is analyzed to output the target predicted path corresponding to the target trailer and other predicted paths corresponding to other traffic participants, including:

[0031] The feature vectors corresponding to the target trailer and other traffic participants are decoded to obtain the predicted displacements of the target trailer and other traffic participants.

[0032] The predicted displacement corresponding to the target trailer is added to the current position information of the target trailer to obtain the target predicted path corresponding to the target trailer.

[0033] The predicted displacements for each other traffic participant are added to the current location information of each other traffic participant to obtain the other predicted paths for each other traffic participant.

[0034] The traffic participant path prediction method provided in this invention decodes the feature vectors corresponding to the target trailer and other traffic participants to obtain the predicted displacements corresponding to the target trailer and other traffic participants, ensuring the accuracy of the obtained predicted displacements. The predicted displacement corresponding to the target trailer is added to the current position information of the target trailer to obtain the target predicted path, ensuring the accuracy of the obtained target predicted path. The predicted displacements corresponding to each other traffic participant are added to the current position information of each other traffic participant to obtain other predicted paths for each other traffic participant, ensuring the accuracy of the obtained other predicted paths for each other traffic participant.

[0035] In conjunction with the first aspect, in the sixth embodiment of the first aspect, obtaining target feature information corresponding to the target trailer and other feature information corresponding to other traffic participants within a preset range includes:

[0036] Acquire driving scenario data within a preset range;

[0037] Identify obstacles in driving scenario data to determine the target trailer and other traffic participants;

[0038] The front and body of the target trailer are marked with boxes to generate target boxes for the body and front of the trailer.

[0039] Add bounding boxes to other traffic participants to generate additional bounding boxes;

[0040] Feature information is extracted from the target bounding boxes of the vehicle body and the front of the vehicle to generate target feature information;

[0041] Extract feature information from other bounding boxes to generate other feature information.

[0042] The traffic participant path prediction method provided in this invention acquires driving scenario data within a preset range; identifies obstacles in the driving scenario data to determine the target trailer and other traffic participants, ensuring the accuracy of the determined target trailer and other traffic participants. The front and body of the target trailer are respectively bounding box labeled to generate target body bounding boxes and target front bounding boxes, ensuring the accuracy of the generated target trailer and target front bounding boxes. This fully considers the shape of the target trailer and the separate characteristics of the front and body of the target trailer. Other traffic participants are bounding box labeled to generate other bounding boxes, ensuring the accuracy of the generated other bounding boxes and fully considering the shapes of other traffic participants. Feature information is extracted from the target body and target front bounding boxes to generate target feature information, ensuring the accuracy of the generated target feature information. Feature information is also extracted from other bounding boxes to generate other feature information, ensuring the accuracy of the generated other feature information.

[0043] According to a second aspect, embodiments of the present invention also provide a traffic participant path prediction device, comprising:

[0044] The acquisition module is used to acquire target feature information corresponding to the target trailer and other feature information corresponding to other traffic participants within a preset range; the target feature information is used to characterize the feature state of the vehicle body target box and the vehicle front target box corresponding to the target trailer, and the other feature information is used to characterize the feature state of other boxes corresponding to other traffic participants.

[0045] The generation module is used to generate a structural diagram of the target trailer and other traffic participants based on the relationship between the target feature information corresponding to the target trailer and other feature information corresponding to other traffic participants.

[0046] The output module is used to input the structure graph into a preset target graph neural network, train the structure graph, and output the target predicted path corresponding to the target trailer and other predicted paths corresponding to other traffic participants.

[0047] The traffic participant path prediction device provided in this embodiment of the invention acquires target feature information corresponding to a target trailer and other feature information corresponding to other traffic participants within a preset range. The target feature information characterizes the feature states of the vehicle body target box and the vehicle front target box corresponding to the target trailer, while the other feature information characterizes the feature states of other boxes corresponding to other traffic participants. This demonstrates that the acquired target feature information corresponding to the target trailer can characterize the shape of the vehicle front and body, and the other feature information corresponding to other traffic participants can also characterize the shape of other traffic participants. This ensures the accuracy of the acquired target feature information and other feature information. Based on the relationship between the target feature information corresponding to the target trailer and other feature information corresponding to other traffic participants, a structure graph corresponding to the target trailer and other traffic participants is generated, ensuring the accuracy of the generated structure graph. Then, the structure graph is input into a preset target graph neural network for training, outputting the target predicted path corresponding to the target trailer and other predicted paths corresponding to other traffic participants, ensuring the accuracy of the predicted target predicted path corresponding to the target trailer and other predicted paths corresponding to other traffic participants. The aforementioned device fully considers the shape of the target trailer and other traffic participants, rather than treating them as point masses. Therefore, it ensures that the predicted path for the target trailer includes not only its x, y, z coordinates but also its shape information. Similarly, the predicted paths for other traffic participants include not only their x, y, z coordinates but also their shape information. This avoids misjudgments due to incomplete information, preventing collisions or cautious stopping situations.

[0048] According to a third aspect, embodiments of the present invention provide an electronic device, including a memory and a processor, which are communicatively connected to each other. The memory stores computer instructions, and the processor executes the computer instructions to perform the traffic participant path prediction method of the first aspect or any embodiment of the first aspect.

[0049] According to a fourth aspect, embodiments of the present invention provide a computer-readable storage medium storing computer instructions for causing a computer to perform the traffic participant path prediction method of the first aspect or any embodiment of the first aspect. Attached Figure Description

[0050] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0051] Figure 1 This is a flowchart of the traffic participant path prediction method provided in the embodiments of the present invention;

[0052] Figure 2 This is a flowchart of a traffic participant path prediction method provided by another embodiment of the present invention;

[0053] Figure 3 This is a flowchart of a traffic participant path prediction method provided by another embodiment of the present invention;

[0054] Figure 4 This is a flowchart of a traffic participant path prediction method provided in another embodiment of the present invention;

[0055] Figure 5 This is a flowchart of a traffic participant path prediction method provided by another embodiment of the present invention;

[0056] Figure 6 This is a functional block diagram of the traffic participant path prediction device provided in the embodiments of the present invention;

[0057] Figure 7 This is a schematic diagram of the hardware structure of the electronic device provided in the embodiments of the present invention. Detailed Implementation

[0058] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0059] It should be noted that the method for predicting traffic participant paths provided in this application can be executed by a device for predicting traffic participant paths. This device can be implemented as part or all of an electronic device through software, hardware, or a combination of both. The electronic device can be a controller in the control system of an intelligent driving vehicle, or it can be a controller independent of the intelligent driving vehicle. The controller can be a server or a terminal. In this application embodiment, the server can be a single server or a server cluster composed of multiple servers. The terminal in this application embodiment can be a smartphone, personal computer, tablet computer, wearable device, or other intelligent hardware device such as an intelligent robot. In the following method embodiments, the execution subject is always described using an electronic device as an example. For ease of description, the tractor unit of a trailer is defined as the cab, and the trailer unit is defined as the body.

[0060] In one embodiment of this application, such as Figure 1 As shown, a method for predicting the path of traffic participants is provided. Taking the application of this method to electronic devices as an example, the method includes the following steps:

[0061] S11. Obtain the target feature information corresponding to the target trailer within the preset range and other feature information corresponding to other traffic participants.

[0062] The preset range can be a preset driving range centered on the target trailer. For example, the preset range can be a driving range within 100 meters of the target trailer or a driving range within 200 meters of the target trailer. Alternatively, the preset range can be a preset driving range centered on a point on the road.

[0063] Specifically, target feature information is used to characterize the feature states of the vehicle body target bounding box and the vehicle front target bounding box corresponding to the target trailer, while other feature information is used to characterize the feature states of other bounding boxes corresponding to other traffic participants. The feature states of the vehicle body target bounding box and the vehicle front target bounding box can include their current position information, length, width, height information, and heading angle information. The feature states of other bounding boxes corresponding to other traffic participants can include their current position information, length, width, height information, and heading angle information.

[0064] Optionally, the electronic device can receive target feature information corresponding to the target trailer within a preset range and other feature information corresponding to other traffic participants sent by other devices or modules, such as the environmental perception module of an autonomous driving operation system. In the autonomous driving system, the environmental perception module is the upstream module for fusion, and information flows from the environmental perception module to the fusion module. It can also receive target feature information corresponding to the target trailer within a preset range and other feature information corresponding to other traffic participants input by the user.

[0065] Optionally, the electronic device can also acquire driving scene images within a preset range, identify the driving scene images, and determine the target feature information corresponding to the target trailer and other feature information corresponding to other traffic participants within the preset range.

[0066] This application does not specifically limit the method by which the electronic device obtains the target feature information corresponding to the target trailer and other feature information corresponding to other traffic participants within a preset range.

[0067] For example, target feature information may include the center coordinates (x1, y1, z1) of the vehicle body target bounding box on the map, the length, width, and height of the vehicle body target bounding box, and the heading angle of the vehicle body target bounding box. Other features may include the center coordinates (x2, y2, z2) of the vehicle front target bounding box on the map, the length, width, and height of the vehicle front target bounding box, the heading angle of the vehicle front target bounding box, the category, the Track ID, and the relative position between the vehicle body target bounding box and the vehicle front target bounding box. Other feature information may include the center coordinates (x1, y1, z1) of other bounding boxes on the map. n ,y n ,z n The dimensions, heading angle, category, and track ID of the other boxes.

[0068] This step will be explained in detail below.

[0069] S12. Based on the relationship between the target feature information corresponding to the target trailer and other feature information corresponding to other traffic participants, generate a structural diagram of the target trailer and other traffic participants.

[0070] Specifically, the electronic device can treat the target trailer and other traffic participants as nodes, with the dimensional information of each node representing its corresponding feature information. That is, the dimensional information corresponding to the target trailer represents its target feature information, and the dimensional information corresponding to other traffic participants represents other feature information. The interaction relationships between each node and other nodes are represented by bidirectional graph edges, generating a structure graph. These interactions can include relative positional relationships, relative speed relationships, relative category relationships, etc., between the target trailer and other traffic participants.

[0071] For example, suppose there are 5 traffic participants in this environment, including one target trailer and 4 other traffic participants, each considered as a node. node1 is the target trailer, with its dimensional information (times, label, id, box1, head1, point1, box2, head2, point2). This dimensional information represents the time sequence, category, tracking number, the length, width, and height of the vehicle's target box, the heading angle of the vehicle's target box, and the x, y, z coordinates of the center position of the vehicle's target box. Similarly, the length, width, and height of the vehicle's front target box, the heading angle of the vehicle's front target box, and the x, y, z coordinates of the center position of the vehicle's front target box are also represented. The interaction relationships between each node and other nodes are represented by bidirectional graph edges. These interactions can be relative positional relationships, relative speed relationships, relative category relationships, etc., between the target trailer and the 4 other traffic participants.

[0072] S13. Input the structure graph into the preset target graph neural network, train the structure graph, and output the target prediction path corresponding to the target trailer and other prediction paths corresponding to other traffic participants.

[0073] Specifically, the electronic device can input the generated structure map into a preset target graph neural network, train the structure map, and output the target predicted path corresponding to the target trailer and other predicted paths corresponding to other traffic participants.

[0074] The electronic device can receive a preset target graph neural network input by the user, and can also receive a preset target graph neural network sent by other devices. The electronic device can also obtain the training dataset corresponding to the preset target graph neural network, train the initial graph neural network, and generate the preset target graph neural network.

[0075] This step will be explained in detail below.

[0076] The traffic participant path prediction method provided in this invention obtains target feature information corresponding to a target trailer and other feature information corresponding to other traffic participants within a preset range. The target feature information characterizes the feature states of the vehicle body target box and the vehicle front target box corresponding to the target trailer, while the other feature information characterizes the feature states of other boxes corresponding to other traffic participants. This demonstrates that the obtained target feature information corresponding to the target trailer can characterize the shape of the target trailer's front and body, and the other feature information corresponding to other traffic participants can also characterize the shape of other traffic participants. This ensures the accuracy of the obtained target feature information and other feature information. Based on the relationship between the target feature information corresponding to the target trailer and other feature information corresponding to other traffic participants, a structure graph corresponding to the target trailer and other traffic participants is generated, ensuring the accuracy of the generated structure graph. Then, the structure graph is input into a preset target graph neural network for training, outputting the target predicted path corresponding to the target trailer and other predicted paths corresponding to other traffic participants, ensuring the accuracy of the predicted target predicted path corresponding to the target trailer and other predicted paths corresponding to other traffic participants. The method described above fully considers the shapes of the target trailer and other traffic participants, without treating them as point masses. Therefore, it ensures that the predicted path for the target trailer includes not only its x, y, z coordinates but also its shape. Similarly, the predicted paths for other traffic participants include not only their x, y, z coordinates but also their shape. This avoids misjudgments due to incomplete information, which could lead to collisions or forced stops.

[0077] In one embodiment of this application, such as Figure 2 As shown, a method for predicting the path of traffic participants is provided. Taking the application of this method to electronic devices as an example, the method includes the following steps:

[0078] S21. Obtain target feature information corresponding to the target trailer within the preset range and other feature information corresponding to other traffic participants.

[0079] Among them, target feature information is used to characterize the feature state of the target vehicle body target box and the target vehicle front target box corresponding to the target trailer, and other feature information is used to characterize the feature state of other boxes corresponding to other traffic participants.

[0080] For details on this step, please refer to [link / reference]. Figure 1 The details of S11 will not be elaborated here.

[0081] S22. Based on the relationship between the target feature information corresponding to the target trailer and other feature information corresponding to other traffic participants, generate a structural diagram of the target trailer and other traffic participants.

[0082] For details on this step, please refer to [link / reference]. Figure 1 The details of S12 will not be elaborated here.

[0083] S23. Input the structure graph into the preset target graph neural network, train the structure graph, and output the target prediction path corresponding to the target trailer and other prediction paths corresponding to other traffic participants.

[0084] S231. Based on the relationships between the target trailer and other traffic participants in the structural diagram, determine the driving priority between the target trailer and other traffic participants.

[0085] Specifically, the electronic device can acquire the pairwise relationships between the target trailer and other traffic participants in the structure diagram. These pairwise relationships can include relationships based on driving speed, type, and current state.

[0086] The electronic equipment determines the driving priority between each pair of the target trailer and other traffic participants based on the relationships between each pair of the target trailer and other traffic participants in the structural diagram.

[0087] This step will be explained in detail below.

[0088] S232. Input the structure map and driving priority into the preset target map neural network, train the structure map and driving priority, and output the target prediction path corresponding to the target trailer and other prediction paths corresponding to other traffic participants.

[0089] Specifically, the electronic device can input the structure map and driving priority into a preset target map neural network. The preset target map neural network encodes and trains the structure map and driving priority, and then decodes the encoded data to output the target predicted path corresponding to the target trailer and other predicted paths corresponding to other traffic participants.

[0090] This step will be explained in detail below.

[0091] The traffic participant path prediction method provided in this invention determines the driving priority between each pair of target trailers and other traffic participants based on the pairwise relationships between them in a structure graph, ensuring the accuracy of the determined driving priority. The structure graph and driving priority are input into a preset target graph neural network for training, outputting the target predicted path for the target trailer and other predicted paths for other traffic participants. This method fully considers the driving priority between each pair of target trailers and other traffic participants, ensuring the accuracy of the output target predicted path and other predicted paths. It avoids misjudgments due to incomplete information, which could lead to collisions or cautious stopping.

[0092] In an optional embodiment of this application, such as Figure 3 As shown, the step S231 above, "determining the driving priority between each pair of the target trailer and other traffic participants based on the pairwise relationships between the target trailer and other traffic participants in the structure diagram," may include the following steps:

[0093] S311. Obtain the driving speed, obstacle type, and current status of the target trailer and other traffic participants.

[0094] The current state is used to characterize the angle between the front and body of the target trailer and other traffic participants, respectively.

[0095] Optionally, the electronic device can acquire driving scene data over a continuous time period transmitted by the sensor device based on the connection between the electronic device and the sensor device, identify the driving scene data, determine the target trailer and other traffic participants, and determine the driving speeds corresponding to the target trailer and other traffic participants based on the identification results. The electronic device can also receive the driving speeds corresponding to the target trailer and other traffic participants input by the user, or it can receive the driving speeds corresponding to the target trailer and other traffic participants transmitted by other devices.

[0096] Optionally, the electronic device can acquire driving scene data over a continuous time period transmitted by the sensor device based on the connection between the electronic device and the sensor device, identify the driving scene data, and determine the obstacle types corresponding to the target trailer and other traffic participants. The obstacle type for the target trailer is a trailer, and the obstacle types for other traffic participants can include motor vehicles, non-motor vehicles, pedestrians, etc., and this application does not specifically limit the obstacle types. The electronic device can also receive user input of the obstacle types corresponding to the target trailer and other traffic participants, and can also receive obstacle types corresponding to the target trailer and other traffic participants transmitted by other devices.

[0097] Optionally, the electronic device can also determine the current status of the target trailer and other traffic participants based on the identified target trailer and other traffic participants.

[0098] Specifically, for the target trailer, its current state is defined by the angle between its cab and body. For other road users, there is no angle between their cab and rear.

[0099] S312. Determine the driving priority between each pair of the target trailer and other traffic participants based on their respective driving speeds, obstacle types, and current states.

[0100] In an optional embodiment of this application, step S312, "determining the driving priority between each pair of the target trailer and other traffic participants based on the relative relationships between their respective driving speeds, obstacle types, and current states," may include the following steps:

[0101] (1) For any two traffic participants among the target trailer and other traffic participants, calculate the speed priority based on the driving speeds of the two traffic participants respectively.

[0102] Specifically, for any two traffic participants among the target trailer and other traffic participants, the electronic device can obtain the driving speeds of the two traffic participants, compare the driving speeds of the two traffic participants, and calculate the speed priority based on the magnitude of the driving speeds of the two traffic participants.

[0103] For example, taking the interaction relationship between node1 and node2 in the structure diagram as an example, the driving priority of node1 over node2 is w12, and the driving priority of node2 over node1 is w21. Initially, both values ​​are 0. For driving speeds of 0-100km / h, the corresponding weights are 0 to 1 respectively, and for speeds exceeding 100km / h, the weight is also set to 1.

[0104] Assume that node1 travels at a speed of 70 km / h and has a weight of 0.7, and node2 travels at a speed of 110 km / h and has a weight of 1.

[0105] (2) Calculate the obstacle category priority based on the obstacle categories corresponding to the two traffic participants.

[0106] Specifically, for any two traffic participants among the target trailer and other traffic participants, the electronic device can obtain the obstacle categories corresponding to the two traffic participants, and then calculate the obstacle category priority based on the obstacle categories corresponding to the two traffic participants.

[0107] For example, the weights for obstacle categories are: cars: 0.2, pedestrians: 0.3, cyclists and motorcycles: 0.4, buses: 0.5, and trailers: 0.6.

[0108] (3) Calculate the priority of the current state based on the current state of the two traffic participants.

[0109] Specifically, for any two traffic participants among the target trailer and other traffic participants, the electronic device can obtain the current state of any two traffic participants and then calculate the priority of the current state based on the current state of any two traffic participants.

[0110] For example, when the included angle is 0 degrees, the weight is 0, and the weight increases by 0.1 for every 10 degrees increase in the included angle.

[0111] (4) Determine the driving priority between the target trailer and other traffic participants based on the speed priority, obstacle category priority and current status priority of any two traffic participants.

[0112] In one optional embodiment of this application, the electronic device can add the speed priority, obstacle category priority, and current state priority of any two traffic participants to calculate the driving priority between the target trailer and other traffic participants.

[0113] In another optional embodiment of this application, the electronic device can obtain the weight information corresponding to speed priority, obstacle category priority and current state priority respectively, multiply the speed priority, obstacle category priority and current state priority by the corresponding weight information respectively, and then add them together to calculate the driving priority between the target trailer and other traffic participants.

[0114] The traffic participant path prediction method provided in this invention obtains the driving speed, obstacle category, and current state of the target trailer and other traffic participants. Then, for any two traffic participants, speed priority is calculated based on their respective driving speeds, ensuring the accuracy of the calculated speed priority. Obstacle category priority is calculated based on the respective obstacle categories of the two traffic participants, ensuring the accuracy of the calculated obstacle category priority. Then, current state priority is calculated based on the respective current states of the two traffic participants, ensuring the accuracy of the calculated current state priority. Based on the speed priority, obstacle category priority, and current state priority of any two traffic participants, the driving priority between each pair of the target trailer and other traffic participants is determined, ensuring the accuracy of the determined driving priority between each pair of the target trailer and other traffic participants.

[0115] In an optional embodiment of this application, such as Figure 4 As shown, the step S232 above, "inputting the structure map and driving priority into a preset target map neural network, training the structure map and driving priority, and outputting the target predicted path corresponding to the target trailer and other predicted paths corresponding to other traffic participants," may include the following steps:

[0116] S321. Input the structure diagram and driving priority into the preset target diagram neural network. The preset target diagram neural network learns and encodes the structure diagram and driving priority to obtain encoded data.

[0117] Specifically, the electronic device can input the structure diagram and driving priority into a preset target graph neural network. The preset target graph neural network extracts features from the structure diagram and driving priority, and learns and encodes based on the extracted features to obtain encoded data.

[0118] S322. Perform graph convolution on the encoded data to generate feature vectors corresponding to the target trailer and other traffic participants within a preset time period in the future.

[0119] Specifically, the electronic device can also perform graph convolution on the generated coded data to generate feature vectors corresponding to the target trailer and other traffic participants within a preset time period in the future.

[0120] S323. Analyze each feature vector and output the target predicted path corresponding to the target trailer and other predicted paths corresponding to other traffic participants.

[0121] In an optional embodiment of this application, the above-mentioned step S323, "analyzing each feature vector and outputting the target predicted path corresponding to the target trailer and other predicted paths corresponding to other traffic participants," may include the following steps:

[0122] (1) Decode the feature vectors corresponding to the target trailer and other traffic participants respectively to obtain the predicted displacements corresponding to the target trailer and other traffic participants respectively.

[0123] Specifically, the electronic device can decode the feature vectors corresponding to the target trailer and other traffic participants to obtain the predicted displacements corresponding to the target trailer and other traffic participants.

[0124] (2) Add the predicted displacement corresponding to the target trailer to the current position information of the target trailer to obtain the target predicted path corresponding to the target trailer.

[0125] Specifically, the electronic device can obtain the current position information of the target trailer, and then add the predicted displacement corresponding to the target trailer to the current position information of the target trailer to obtain the target predicted path corresponding to the target trailer.

[0126] (3) Add the predicted displacements of each other traffic participant to the current location information of each other traffic participant to obtain the other predicted paths of each other traffic participant.

[0127] Specifically, the electronic device can obtain the current location information of each other traffic participant, and then add the predicted displacement corresponding to each other traffic participant to the current location information of each other traffic participant to obtain the other predicted paths corresponding to each other traffic participant.

[0128] The traffic participant path prediction method provided in this invention inputs a structure graph and driving priorities into a preset target graph neural network. The preset target graph neural network learns and encodes the structure graph and driving priorities to obtain encoded data, ensuring the accuracy of the obtained encoded data. Graph convolution is performed on the encoded data to generate feature vectors corresponding to the target trailer and other traffic participants within a preset future time period, ensuring the accuracy of the generated feature vectors corresponding to the target trailer and other traffic participants within the preset future time period. Decoding is performed on the feature vectors corresponding to the target trailer and other traffic participants to obtain the predicted displacements corresponding to the target trailer and other traffic participants respectively, ensuring the accuracy of the obtained predicted displacements. The predicted displacement corresponding to the target trailer is added to the current position information of the target trailer to obtain the target predicted path corresponding to the target trailer, ensuring the accuracy of the obtained target predicted path corresponding to the target trailer. The predicted displacements corresponding to each other traffic participant are added to the current position information of each other traffic participant to obtain other predicted paths corresponding to each other traffic participant, ensuring the accuracy of the obtained other predicted paths corresponding to other traffic participants.

[0129] In one embodiment of this application, such as Figure 5 As shown, a method for predicting the path of traffic participants is provided. Taking the application of this method to electronic devices as an example, the method includes the following steps:

[0130] S41. Obtain target feature information corresponding to the target trailer within the preset range and other feature information corresponding to other traffic participants.

[0131] Among them, target feature information is used to characterize the feature state of the target vehicle body target box and the target vehicle front target box corresponding to the target trailer, and other feature information is used to characterize the feature state of other boxes corresponding to other traffic participants.

[0132] In an optional embodiment of this application, step S41, "obtaining target feature information corresponding to the target trailer and other feature information corresponding to other traffic participants within a preset range," may include the following steps:

[0133] S411. Obtain driving scenario data within a preset range.

[0134] Optionally, the electronic device can acquire driving scene data within a preset range sent by the sensor device based on the connection between the electronic device and the sensor device. The sensor device can be at least one of a camera device, a lidar device, a millimeter-wave radar device, or similar sensor devices.

[0135] S412. Identify obstacles in driving scenario data to determine the target trailer and other traffic participants.

[0136] Specifically, electronic devices can use obstacle recognition models to identify obstacles in driving scenario data and determine the target trailer and other traffic participants.

[0137] The obstacle recognition model can be user input received by the electronic device, or it can be sent by other devices. The obstacle recognition model can also be trained by the electronic device based on obstacle recognition training data.

[0138] S413. Mark the front and body of the target trailer with boxes to generate target boxes for the body and front.

[0139] Specifically, after identifying the target trailer and other traffic participants, the electronic device can mark the front and body of the target trailer with bounding boxes, generating target bounding boxes for the body and front of the trailer.

[0140] S414. Mark other traffic participants with boxes to generate other boxes.

[0141] Specifically, after identifying the target trailer and other traffic participants, the electronic device can mark other traffic participants with bounding boxes and generate other bounding boxes.

[0142] S415. Extract feature information from the vehicle body target bounding box and the vehicle front target bounding box to generate target feature information.

[0143] Specifically, after the electronic device marks the target trailer and other traffic participants with bounding boxes, it can extract feature information from the vehicle body target bounding box and the vehicle front target bounding box to generate target feature information.

[0144] S416. Extract feature information from other boxes to generate other feature information.

[0145] Specifically, after the target trailer and other traffic participants are bounded by boxes, the electronic device can extract feature information from the other boxes and generate other feature information.

[0146] S42. Based on the relationship between the target feature information corresponding to the target trailer and other feature information corresponding to other traffic participants, generate a structural diagram of the target trailer and other traffic participants.

[0147] For details on this step, please refer to [link / reference]. Figure 2 The details of S22 will not be elaborated here.

[0148] S43. Input the structure graph into the preset target graph neural network, train the structure graph, and output the target prediction path corresponding to the target trailer and other prediction paths corresponding to other traffic participants.

[0149] For details on this step, please refer to [link / reference]. Figure 2 The details of S23 will not be elaborated here.

[0150] The traffic participant path prediction method provided in this invention acquires driving scenario data within a preset range; identifies obstacles in the driving scenario data to determine the target trailer and other traffic participants, ensuring the accuracy of the determined target trailer and other traffic participants. The front and body of the target trailer are respectively bounding box labeled to generate target body bounding boxes and target front bounding boxes, ensuring the accuracy of the generated target trailer and target front bounding boxes. This fully considers the shape of the target trailer and the separate characteristics of the front and body of the target trailer. Other traffic participants are bounding box labeled to generate other bounding boxes, ensuring the accuracy of the generated other bounding boxes and fully considering the shapes of other traffic participants. Feature information is extracted from the target body and target front bounding boxes to generate target feature information, ensuring the accuracy of the generated target feature information. Feature information is also extracted from other bounding boxes to generate other feature information, ensuring the accuracy of the generated other feature information.

[0151] It should be understood that, although Figures 1-5 The steps in the flowchart are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order in which these steps are executed, and they can be performed in other orders. Figures 1-5 At least some of the steps in the process may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but may be executed at different times. The execution order of these steps or stages is not necessarily sequential, but may be executed in turn or alternately with other steps or at least some of the steps or stages in other steps.

[0152] like Figure 6 As shown, this embodiment provides a traffic participant path prediction device, including:

[0153] The acquisition module 51 is used to acquire target feature information corresponding to the target trailer and other feature information corresponding to other traffic participants within a preset range; the target feature information is used to characterize the feature state of the vehicle body target box and the vehicle front target box corresponding to the target trailer, and the other feature information is used to characterize the feature state of other boxes corresponding to other traffic participants.

[0154] The generation module 52 is used to generate a structural diagram of the target trailer and other traffic participants based on the relationship between the target feature information corresponding to the target trailer and other feature information corresponding to other traffic participants.

[0155] The output module 53 is used to input the structure graph into the preset target graph neural network, train the structure graph, and output the target prediction path corresponding to the target trailer and other prediction paths corresponding to other traffic participants.

[0156] In one embodiment of this application, the output module 53 is specifically used to determine the driving priority between the target trailer and other traffic participants based on the pairwise relationships between the target trailer and other traffic participants in the structure graph; input the structure graph and driving priority into a preset target graph neural network, train the structure graph and driving priority, and output the target prediction path corresponding to the target trailer and other prediction paths corresponding to other traffic participants.

[0157] In one embodiment of this application, the output module 53 is specifically used to obtain the driving speed, obstacle type, and current state of the target trailer and other traffic participants respectively; wherein, the current state is used to characterize the angle between the front and body of the target trailer and other traffic participants respectively; and to determine the driving priority between each pair of the target trailer and other traffic participants based on the relative relationship between the driving speed, obstacle type, and current state of the target trailer and other traffic participants respectively.

[0158] In one embodiment of this application, the output module 53 is specifically used to calculate speed priority for any two traffic participants among the target trailer and other traffic participants, based on their respective driving speeds; calculate obstacle category priority based on the respective obstacle categories of the two traffic participants; calculate current state priority based on the respective current states of the two traffic participants; and determine the driving priority between any two traffic participants based on their speed priority, obstacle category priority, and current state priority.

[0159] In one embodiment of this application, the output module 53 is specifically used to input the structure graph and driving priority into a preset target graph neural network. The preset target graph neural network learns and encodes the structure graph and driving priority to obtain encoded data. The encoded data is then subjected to graph convolution to generate feature vectors corresponding to the target trailer and other traffic participants within a preset time period in the future. Each feature vector is analyzed to output the target predicted path corresponding to the target trailer and other predicted paths corresponding to other traffic participants.

[0160] In one embodiment of this application, the output module 53 is specifically used to decode the feature vectors corresponding to the target trailer and other traffic participants to obtain the predicted displacements corresponding to the target trailer and other traffic participants respectively; add the predicted displacement corresponding to the target trailer to the current position information of the target trailer to obtain the target predicted path corresponding to the target trailer; and add the predicted displacements corresponding to each other traffic participant to the current position information of each other traffic participant to obtain the other predicted paths corresponding to each other traffic participant.

[0161] In one embodiment of this application, the acquisition module 51 is specifically used to acquire driving scene data within a preset range; identify obstacles in the driving scene data to determine the target trailer and other traffic participants; mark the front and body of the target trailer with bounding boxes to generate target bounding boxes for the vehicle body and the front; mark other traffic participants with bounding boxes to generate other bounding boxes; extract feature information from the target bounding boxes for the vehicle body and the front; and extract feature information from the other bounding boxes to generate other feature information.

[0162] For specific limitations and beneficial effects regarding the traffic participant route prediction device, please refer to the limitations of the traffic participant route prediction method above, which will not be repeated here. Each module in the aforementioned traffic participant route prediction device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in the electronic device in hardware form, or stored in the memory of the electronic device in software form, so that the processor can call and execute the corresponding operations of each module.

[0163] This invention also provides an electronic device having the above-described features. Figure 6 The traffic participant path prediction device shown.

[0164] like Figure 7 As shown, Figure 7 This is a schematic diagram of the structure of an electronic device provided in an optional embodiment of the present invention, such as... Figure 7As shown, the electronic device may include: at least one processor 61, such as a CPU (Central Processing Unit), at least one communication interface 63, memory 64, and at least one communication bus 62. The communication bus 62 is used to enable communication between these components. The communication interface 63 may include a display screen or a keyboard; optionally, the communication interface 63 may also include a standard wired interface or a wireless interface. The memory 64 may be high-speed RAM (Random Access Memory) or non-volatile memory, such as at least one disk storage device. Optionally, the memory 64 may also be at least one storage device located remotely from the aforementioned processor 61. The processor 61 may be combined with... Figure 6 The described apparatus has an application program stored in memory 64, and the processor 61 calls the program code stored in memory 64 to perform any of the above method steps.

[0165] The communication bus 62 can be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The communication bus 62 can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 7 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0166] The memory 64 may include volatile memory, such as random-access memory (RAM); the memory may also include non-volatile memory, such as flash memory, hard disk drive (HDD) or solid-state drive (SSD); the memory 64 may also include a combination of the above types of memory.

[0167] The processor 61 can be a central processing unit (CPU), a network processor (NP), or a combination of CPU and NP.

[0168] The processor 61 may further include a hardware chip. This hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The PLD may be a complex programmable logic device (CPLD), a field-programmable gate array (FPGA), a generic array logic (GAL), or any combination thereof.

[0169] Optionally, memory 64 is also used to store program instructions. Processor 61 can invoke program instructions to implement the functions described in this application. Figures 1 to 5 The traffic participant path prediction method shown in the embodiment.

[0170] This invention also provides a non-transitory computer storage medium storing computer-executable instructions that can execute the traffic participant path prediction method in any of the above method embodiments. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk drive (HDD), or solid-state drive (SSD), etc.; the storage medium may also include combinations of the above types of memory.

[0171] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A method for predicting the routes of traffic participants, characterized in that, include: Acquire target feature information corresponding to the target trailer within a preset range and other feature information corresponding to other traffic participants; The target feature information is used to characterize the feature state of the vehicle body target box and the vehicle front target box corresponding to the target trailer, and the other feature information is used to characterize the feature state of the other boxes corresponding to the other traffic participants; Based on the relationship between the target feature information corresponding to the target trailer and the other feature information corresponding to the other traffic participants, a structural diagram corresponding to the target trailer and the other traffic participants is generated; The structure graph is input into a preset target graph neural network to train the structure graph and output the target prediction path corresponding to the target trailer and other prediction paths corresponding to the other traffic participants. The step of inputting the structure graph into a preset target graph neural network, training the structure graph, and outputting the target predicted path corresponding to the target trailer and other predicted paths corresponding to other traffic participants includes: Based on the relationships between the target trailer and other traffic participants in the structural diagram, determine the driving priority between each pair of the target trailer and other traffic participants; The structure diagram and the driving priority are input into a preset target graph neural network to train the structure diagram and the driving priority, and output the target prediction path corresponding to the target trailer and other prediction paths corresponding to the other traffic participants. The step of inputting the structure diagram and the driving priority into a preset target graph neural network, training the structure diagram and the driving priority, and outputting the target predicted path corresponding to the target trailer and other predicted paths corresponding to the other traffic participants includes: The structure diagram and the driving priority are input into a preset target graph neural network, and the preset target graph neural network learns and encodes the structure diagram and the driving priority to obtain encoded data; Graph convolution is performed on the encoded data to generate feature vectors corresponding to the target trailer and other traffic participants within a preset future time period; The feature vectors are analyzed to output the target predicted path corresponding to the target trailer and other predicted paths corresponding to the other traffic participants.

2. The method according to claim 1, characterized in that, The step of determining the driving priority between each pair of the target trailer and other traffic participants based on the pairwise relationships between them in the structural diagram includes: The driving speed, obstacle type, and current state of the target trailer and other traffic participants are obtained respectively; wherein, the current state is used to characterize the angle between the front and body of the target trailer and other traffic participants respectively. Based on the relative relationships between the target trailer and the other traffic participants' respective driving speeds, obstacle types, and current states, the driving priority between each pair of the target trailer and the other traffic participants is determined.

3. The method according to claim 2, characterized in that, The step of determining the driving priority between each pair of the target trailer and other traffic participants based on the relative relationships between their respective driving speeds, obstacle types, and current states includes: For any two traffic participants among the target trailer and other traffic participants, calculate the speed priority based on the respective driving speeds of the two traffic participants. Based on the obstacle categories corresponding to the two traffic participants, calculate the obstacle category priority; Calculate the priority of the current state based on the current state of the two traffic participants; Based on the speed priority, obstacle category priority, and current state priority of any two traffic participants, the driving priority between the target trailer and the other traffic participants is determined.

4. The method according to claim 1, characterized in that, The step of analyzing the feature vector and outputting the target predicted path corresponding to the target trailer and other predicted paths corresponding to other traffic participants includes: The feature vectors corresponding to the target trailer and the other traffic participants are decoded to obtain the predicted displacements corresponding to the target trailer and the other traffic participants. The predicted displacement corresponding to the target trailer is added to the current position information of the target trailer to obtain the target predicted path corresponding to the target trailer. The predicted displacement corresponding to each of the other traffic participants is added to the current location information of each of the other traffic participants to obtain the other predicted path corresponding to each of the other traffic participants.

5. The method according to claim 1, characterized in that, The acquisition of target feature information corresponding to the target trailer and other feature information corresponding to other traffic participants within the preset range includes: Acquire driving scenario data within a preset range; The obstacles in the driving scenario data are identified to determine the target trailer and other traffic participants; The front and body of the target trailer are labeled with boxes to generate the target boxes for the body and the target boxes for the front. The other traffic participants are labeled with boxes to generate the other boxes; Feature information is extracted from the vehicle body target bounding box and the vehicle front target bounding box to generate the target feature information; Feature information is extracted from the other bounding boxes to generate the other feature information.

6. A traffic participant route prediction device, characterized in that, include: The acquisition module is used to acquire target feature information corresponding to the target trailer and other feature information corresponding to other traffic participants within a preset range; The target feature information is used to characterize the feature state of the vehicle body target box and the vehicle front target box corresponding to the target trailer, and the other feature information is used to characterize the feature state of the other boxes corresponding to the other traffic participants; The generation module is used to generate a structural diagram corresponding to the target trailer and other traffic participants based on the relationship between the target feature information corresponding to the target trailer and the other feature information corresponding to the other traffic participants; The output module is used to input the structure graph into a preset target graph neural network, train the structure graph, and output the target predicted path corresponding to the target trailer and other predicted paths corresponding to the other traffic participants; wherein, the step of inputting the structure graph into the preset target graph neural network, training the structure graph, and outputting the target predicted path corresponding to the target trailer and other predicted paths corresponding to the other traffic participants includes: determining the driving priority between each pair of the target trailer and the other traffic participants based on the pairwise relationships between the target trailer and the other traffic participants in the structure graph; inputting the structure graph and the driving priority into the preset target graph neural network, training the structure graph and the driving priority, and outputting the target predicted path corresponding to the target trailer. And other predicted paths corresponding to the other traffic participants; wherein, the step of inputting the structure graph and the driving priority into a preset target graph neural network, training the structure graph and the driving priority, and outputting the target predicted path corresponding to the target trailer and other predicted paths corresponding to the other traffic participants includes: inputting the structure graph and the driving priority into the preset target graph neural network, the preset target graph neural network learning and encoding the structure graph and the driving priority to obtain encoded data; performing graph convolution on the encoded data to generate feature vectors corresponding to the target trailer and other traffic participants within a preset future time period; analyzing each of the feature vectors, and outputting the target predicted path corresponding to the target trailer and other predicted paths corresponding to the other traffic participants.

7. An electronic device, characterized in that, The method includes a memory and a processor, wherein the memory stores computer instructions, and the processor executes the computer instructions to perform the traffic participant path prediction method according to any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing a computer to perform the traffic participant path prediction method according to any one of claims 1-5.

Images