A vehicle trajectory planning method, device, equipment and storage medium

Abstract

The application discloses a vehicle trajectory planning method and device, equipment and storage medium, relates to the path planning technical field, is applied to the current vehicle driving on the main road, and includes the following steps: collecting vehicle feature information of each target vehicle located in the merging area in the vehicle merging scene; the merging area includes the main road and the ramp area; based on the road feature information of the vehicle merging scene and the vehicle feature information, the merging intention of each target vehicle is predicted by using a preset neural network, and each prediction result is obtained; the preset neural network is a neural network constructed based on a three-layer bidirectional long short-term memory network; the driving trajectory of the current vehicle is planned according to each prediction result, and a target driving trajectory for avoiding each target vehicle is obtained. In this way, in the application, the vehicles on the main road can predict the intention of each target vehicle on the ramp to merge into the main road in the vehicle merging scene in advance, and the driving trajectory that can avoid each target vehicle can be planned, thereby improving the driving safety of the vehicle.

Description

Technical Field

[0001] This invention relates to the field of path planning technology, and in particular to a vehicle trajectory planning method, apparatus, device, and storage medium. Background Technology

[0002] When a vehicle passes through a main road with a merging area, it is often affected by vehicles merging into the main road from the ramp when it is traveling in the main road lane close to the merging ramp. In this case, it will cause a large lateral interference to the driving status of vehicles on the main road. Direct trajectory prediction is often ineffective at this time, which will have a significant impact on the vehicle's lateral and longitudinal planning and control modules, and ultimately a significant impact on the vehicle's safety.

[0003] It can be seen that how to perform path planning for vehicles traveling on the main road in the scenario of vehicle merging, so as to eliminate the safety hazards caused by vehicles merging into the main road from the ramp, is a problem to be solved in this field. Summary of the Invention

[0004] In view of this, the purpose of this invention is to provide a vehicle trajectory planning method, apparatus, device, and storage medium. Vehicles traveling on the main road can predict the intentions of target vehicles located on ramps to merge into the main road in advance during vehicle merging scenarios, enabling more accurate planning of driving trajectories that avoid these target vehicles and improving vehicle driving safety. The specific solution is as follows:

[0005] Firstly, this application provides a vehicle trajectory planning method, applied to a vehicle currently traveling on a main road, including:

[0006] In a vehicle merging scenario, vehicle characteristic information of each target vehicle located in the merging area is collected; the merging area includes the main road and the ramp area that meets the preset distance condition from the main road;

[0007] Based on the road feature information of the vehicle merging scenario and the feature information of each vehicle, a preset neural network is used to predict the merging intention of each target vehicle and obtain the corresponding prediction results; the preset neural network is a neural network constructed based on a three-layer bidirectional long short-term memory network.

[0008] Based on the prediction results, the current vehicle's trajectory is planned to obtain a target trajectory for avoiding the target vehicles when the vehicles merge.

[0009] Optionally, before collecting vehicle feature information of each target vehicle located in the merging area in the vehicle merging scenario, the method further includes:

[0010] Based on the surrounding map data of the current vehicle, determine whether the current driving scene of the vehicle matches the preset merging scene;

[0011] If the current vehicle's driving scenario matches the preset merging scenario, a topology map is drawn based on the surrounding map data to obtain a merging area topology map, so as to collect vehicle feature information of each target vehicle located in the merging area based on the merging area topology map in the vehicle merging scenario.

[0012] Optionally, the step of collecting vehicle feature information of each target vehicle located in the merging area in a vehicle merging scenario includes:

[0013] In a vehicle merging scenario, attribute values ​​of each target vehicle located in the merging area are collected by cameras and / or radar, and the attribute values ​​are marked on the topology map of the merging area to obtain vehicle feature information of each target vehicle.

[0014] Optionally, the step of predicting the merging intention of each target vehicle based on the road feature information of the vehicle merging scenario and the feature information of each vehicle, to obtain corresponding prediction results, includes:

[0015] The road feature information of the vehicle merging scene is processed by the first bidirectional long short-term memory network to obtain the corresponding map tensor.

[0016] The feature information of each vehicle is extracted using a second bidirectional long short-term memory network to obtain the corresponding target tensor.

[0017] By using a third bidirectional long short-term memory network and the map tensor and the target tensor, the merging intention of each target vehicle is predicted, and the corresponding prediction results are obtained.

[0018] Optionally, the prediction result characterizes the probability and corresponding location of the target vehicle merging from the ramp area into the main road at each time point.

[0019] Optionally, the step of planning the current vehicle's trajectory based on each of the prediction results includes:

[0020] Based on the prediction results, obstacle information is marked at the location points corresponding to the target vehicles at each time point, so as to plan the current vehicle's driving trajectory based on the obstacle information.

[0021] Optionally, the step of planning the current vehicle's trajectory based on the prediction results to obtain a target trajectory for avoiding the target vehicles in the vehicle merging scenario includes:

[0022] Based on the prediction results, road feature information, and vehicle feature information, the current vehicle's driving trajectory is planned to obtain a target driving trajectory for avoiding the target vehicles in the vehicle merging scenario.

[0023] Secondly, this application provides a vehicle trajectory planning device, applied to a vehicle currently traveling on a main road, comprising:

[0024] The information acquisition module is used to collect vehicle characteristic information of each target vehicle located in the merging area in a vehicle merging scenario; the merging area includes the main road and the ramp area that meets the preset distance condition from the main road;

[0025] The prediction module is used to predict the merging intention of each target vehicle based on the road feature information of the vehicle merging scene and the feature information of each vehicle, and to obtain the corresponding prediction results; the preset neural network is a neural network constructed based on a three-layer bidirectional long short-term memory network.

[0026] The planning module is used to plan the driving trajectory of the current vehicle based on the prediction results, so as to obtain a target driving trajectory for avoiding the target vehicles when the vehicles merge into the scene.

[0027] Thirdly, this application provides an electronic device, comprising:

[0028] Memory, used to store computer programs;

[0029] A processor is used to execute the computer program to implement the vehicle trajectory planning method described above.

[0030] Fourthly, this application provides a computer-readable storage medium for storing a computer program that, when executed by a processor, implements the vehicle trajectory planning method described above.

[0031] Therefore, in this application, a vehicle traveling on the main road can collect vehicle feature information of each target vehicle located in the merging area during a vehicle merging scenario. The merging area includes the main road and ramp areas that meet preset distance conditions from the main road. Then, based on the road feature information of the vehicle merging scenario and the feature information of each vehicle, the merging intention of each target vehicle is predicted, and corresponding prediction results are obtained. Furthermore, the driving trajectory of the current vehicle is planned according to the prediction results to obtain a target driving trajectory for avoiding each target vehicle in the vehicle merging scenario. In this way, in this application, a vehicle traveling on the main road can predict the intention of each target vehicle located on the ramp to merge into the main road in advance during a vehicle merging scenario, and obtain corresponding prediction results. Then, by combining these prediction results, a driving trajectory that can avoid each target vehicle can be planned more accurately, improving vehicle driving safety in vehicle merging scenarios. Attached Figure Description

[0032] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0033] Figure 1 This is a flowchart of a vehicle trajectory planning method disclosed in this application;

[0034] Figure 2 This is a schematic diagram of a vehicle merging scenario disclosed in this application;

[0035] Figure 3 This is a schematic diagram of a network framework disclosed in this application;

[0036] Figure 4 This is a flowchart of a specific vehicle trajectory planning method disclosed in this application;

[0037] Figure 5 This is a schematic diagram of a specific network structure disclosed in this application;

[0038] Figure 6 This is a schematic diagram of the structure of a vehicle trajectory planning device disclosed in this application;

[0039] Figure 7 This is a structural diagram of an electronic device disclosed in this application. Detailed Implementation

[0040] 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, and 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.

[0041] See Figure 1 As shown, this embodiment of the invention discloses a vehicle trajectory planning method, applied to a vehicle currently traveling on a main road, including:

[0042] Step S11: Collect vehicle feature information of each target vehicle located in the merging area in the vehicle merging scenario; the merging area includes the main road and the ramp area that meets the preset distance condition from the main road.

[0043] In this embodiment, in the vehicle merging scenario, the merging area may include the main road and the connected ramp area. The target vehicle located in the ramp area may merge into the main road at any time. In the vehicle merging scenario, the current vehicle traveling on the main road can collect the vehicle feature information of each target vehicle in the merging area so as to perform trajectory planning for the current vehicle based on the vehicle feature information of each target vehicle.

[0044] In one specific embodiment, before collecting vehicle feature information of each target vehicle located in the merging area in a vehicle merging scenario, the method may further include: determining whether the current vehicle's driving scenario matches a preset merging scenario based on the surrounding map data; if the current vehicle's driving scenario matches the preset merging scenario, then drawing a topology map based on the surrounding map data to obtain a merging area topology map, so as to collect vehicle feature information of each target vehicle located in the merging area in the vehicle merging scenario based on the merging area topology map. Specifically, when the current vehicle is driving on the main road, it can be determined whether the current vehicle's driving scenario matches a preset merging scenario based on the surrounding map data; the vehicle merging scenario can be as follows: Figure 2 As shown, the merging area in the vehicle merging scenario can be divided into three parts. Figure 2The road within the green box is the main road, the road within the red box is the ramp, and the unboxed areas are acceleration and deceleration lanes. It's understandable that if the current vehicle's current driving scenario matches the aforementioned vehicle merging scenario, a topology map of the merging area can be drawn based on surrounding map data. Then, the attribute values ​​of each target vehicle in this merging area can be collected, and these attribute values ​​can be labeled on the topology map to obtain the vehicle characteristic information of each target vehicle in the ramp area. In a specific embodiment, the attribute values ​​of target vehicles in other lanes of the main road can also be considered, and the lane-changing intentions of related vehicles merging from their lanes into the current vehicle's lane can be predicted to plan the current vehicle's trajectory and avoid dangerous situations.

[0045] In another specific embodiment, the step of collecting vehicle feature information of each target vehicle located in the merging area during a vehicle merging scenario may include: collecting vehicle feature information of each target vehicle located in the merging area using a camera and / or radar during the merging scenario. Specifically, the current vehicle can collect attribute values ​​such as speed, acceleration, vehicle type, and heading angle of each target vehicle located in the merging area through a camera or radar, and finally obtain the vehicle feature information of each target vehicle.

[0046] Step S12: Based on the road feature information of the vehicle merging scene and the feature information of each vehicle, a preset neural network is used to predict the merging intention of each target vehicle and obtain the corresponding prediction results; the preset neural network is a neural network constructed based on a three-layer bidirectional long short-term memory network.

[0047] In this embodiment, the vehicle feature information of each target vehicle in the vehicle merging scenario can be obtained through the above steps. Then, based on the road feature information corresponding to the vehicle merging scenario and the related vehicle feature information, the merging intention of each target vehicle from the ramp area to the main road can be predicted. A neural network is pre-constructed based on a three-layer long short-term memory network. The prediction result of each target vehicle merging from the ramp to the main road can be predicted through the preset neural network.

[0048] In one specific embodiment, the step of predicting the merging intention of each target vehicle using a preset neural network based on the road feature information of the vehicle merging scene and the feature information of each vehicle to obtain corresponding prediction results may include: performing feature extraction processing on the road feature information of the vehicle merging scene through a first bidirectional long short-term memory network to obtain a corresponding map tensor; performing feature extraction processing on the feature information of each vehicle through a second bidirectional long short-term memory network to obtain a corresponding target tensor; and performing merging intention prediction on each target vehicle using a third bidirectional long short-term memory network and the map tensor and the target tensor to obtain corresponding prediction results. Specifically, in the process of predicting the merging intention of each target vehicle, three bidirectional long short-term memory (BLS) networks are used to extract features for intention prediction. The first BLS network extracts road features, processing the road feature information of the vehicle merging scene. The input information mainly includes the road topology, and the output is the corresponding map tensor. The second BLS network extracts the features of the target vehicles. The input information mainly includes the target vehicle's geodetic coordinates, lateral and longitudinal velocities, lateral and longitudinal accelerations, vehicle type, length, width, height, and heading angle, and the final output is the target tensor corresponding to the target vehicle. The third BLS network predicts the intention of each target vehicle based on the map tensor and target tensor obtained from the first two BLS networks, predicting the intention of each target vehicle to merge into the main road from the corresponding ramp area, and obtaining the corresponding prediction results. The framework of the three-layer bidirectional long short-term memory (BLS) network is as follows: Figure 3 As shown. In a specific embodiment, the prediction result characterizes the probability and corresponding location of the target vehicle merging from the ramp area into the main road at each time point.

[0049] Step S13: Based on the prediction results, plan the driving trajectory of the current vehicle to obtain a target driving trajectory for avoiding each of the target vehicles in the vehicle merging scenario.

[0050] In this embodiment, the above steps can be used to obtain the prediction results of each target vehicle in the vehicle merging scenario that will merge into the main road from the ramp area. Then, the current vehicle's driving trajectory can be planned based on the prediction results, so that the current vehicle can avoid each target vehicle when executing the planned driving trajectory, thus ensuring vehicle driving safety.

[0051] In one specific embodiment, planning the current vehicle's trajectory based on the prediction results may include: marking obstacle information at the location points corresponding to each target vehicle at each time point based on the prediction results, so as to plan the current vehicle's trajectory based on the obstacle information. Specifically, the prediction results for each target vehicle represent the probability and corresponding location point of the target vehicle merging into the main road from the ramp area at each time point. Obstacle information is marked at the location points corresponding to each target vehicle at each time point, and the current vehicle's trajectory can be planned based on this obstacle information.

[0052] In another specific embodiment, the step of planning the current vehicle's trajectory based on the prediction results to obtain a target trajectory for avoiding the target vehicles in the vehicle merging scenario can include: planning the current vehicle's trajectory based on the prediction results, the road feature information, and the vehicle feature information to obtain a target trajectory for avoiding the target vehicles in the vehicle merging scenario. Specifically, during trajectory planning, the prediction results of the target vehicles, the road feature information of the vehicle merging scenario, and the feature information of each vehicle can be considered, and the current vehicle's trajectory can be planned using a pre-set path planning algorithm; the path planning algorithm used can be the A* algorithm, which combines the prediction results corresponding to each target vehicle to plan the current vehicle's trajectory, resulting in a suitable and safe trajectory that can avoid the vehicles.

[0053] Therefore, in this application, vehicles traveling on the main road can predict the intentions of each target vehicle located on the ramp to merge into the main road in advance in the vehicle merging scenario. By using a three-layer long short-term memory network, the merging intentions of each target vehicle in the merging area can be predicted to obtain the corresponding prediction results. Then, by combining the intention prediction results, the driving trajectory that can avoid each target vehicle can be planned more accurately, thereby improving the driving safety of vehicles in the vehicle merging scenario.

[0054] like Figure 4 As shown in the figure, this application discloses a vehicle trajectory planning method, applied to a vehicle currently traveling on a main road, specifically including:

[0055] In this embodiment, a vehicle traveling on the main road can determine whether its current driving scenario is a merging scenario based on high-precision map data of the surrounding environment. If it is determined that the current moment meets the criteria for a vehicle merging scenario, the surrounding scene information can be extracted to draw a topology map of the merging area. Then, target vehicles located in the ramp area identified by sensing camera data or radar data are assigned to the merging area topology map. Subsequently, the trajectory of each target vehicle can be predicted based on the topology map, specifically analyzing target vehicles that conflict with the future planned trajectory of the current vehicle (autonomous vehicle), and adjusting the vehicle's trajectory accordingly. The determination of a vehicle merging scenario can be based on positioning data and high-precision map data, primarily relying on the presence of a merging (connection area between the main road and the ramp) attribute and the distance to the merging point. It is understood that the planned trajectory of the autonomous vehicle can include x, y, z, heading angle, speed, and acceleration information of trajectory points generated every 0.1 seconds based on the vehicle's coordinate system. The high-precision map data includes merging and diverging information, lane line information, lane width, and speed limit information. The positioning data includes the latitude and longitude coordinates of the autonomous vehicle. Perception camera data or radar data refers to the detected attribute values ​​of the target vehicle, including the target vehicle's x, y, z coordinates, speed, acceleration, vehicle type, and heading angle information generated based on the vehicle's own coordinate system. Furthermore, when assigning target vehicles on the topology map, the assignment is based on the vehicle's x, y, z coordinates and heading angle, only assigning vehicles that pose a spatial and temporal collision risk to the vehicle on the topology map. For example, the distances of the vehicle and the target vehicle from the merging point are divided by their respective speeds to obtain t1 and t2. If the absolute value of the difference between t1 and t2 is less than 10 seconds, a collision risk is considered. Correspondingly, when predicting the merging intention of a target vehicle, the merging intention within 5 seconds can be predicted (or intention prediction only begins when the vehicle is in the acceleration / deceleration lane of the merging area). Intentions include two types: intention to merge onto the main road and intention not to merge onto the main road.

[0056] In a specific embodiment, when performing intent prediction using a three-layer bidirectional long short-term memory network, such as Figure 5As shown, A participates in the forward computation, and A' participates in the backward computation. The final output value y depends on the sum of A and A'. Several y values ​​represent the predicted position of the target in each frame. Based on these y values, the target intent is statistically determined, resulting in a more accurate prediction. Here, A represents the forward prediction result, and A' represents the backward prediction result; X refers to the input of each frame, including the map tensor and the target tensor; y is the predicted position point in each frame; S represents the hidden layer for the forward computation, and S' represents the hidden layer for the backward computation. Further, the formulas for calculating A and A', and for calculating y, can be: y² = g(VA² + V'A'²); A'² = f(W'A'³ + U'x²); A² = f(WA1 + Ux²); where W represents the weight parameters for the forward computation, W' represents the weight parameters for the backward computation, U is the cell state for the forward computation, and U' is the cell state for the backward computation. f and g represent the model parameters, V represents the forward concatenation weights, and V' represents the backward concatenation weights. Furthermore, after predicting the trajectories of each target vehicle through the above steps, the A* algorithm can be used to plan the trajectory of the vehicle based on the prediction results. The distance between the vehicle and the target vehicle with the highest probability of merging at each time point (0.1-second interval) is calculated, and obstacle information is assigned at that position, so that the A* algorithm can plan a suitable driving trajectory for the vehicle that can avoid each target vehicle in the merging scenario.

[0057] Therefore, in this application, vehicles traveling on the main road can predict the merging intentions of target vehicles located on ramps in advance during vehicle merging scenarios. A three-layer long short-term memory network can be used to predict the merging intentions of each target vehicle in the merging area, obtaining the corresponding prediction results. Then, combining these intention prediction results allows for more accurate planning of driving trajectories that avoid the target vehicles. This low-cost, low-computing-power trajectory planning improves vehicle driving safety in merging scenarios.

[0058] like Figure 6 As shown in the figure, this application discloses a vehicle trajectory planning device, applied to a vehicle currently traveling on a main road, including:

[0059] The information acquisition module 11 is used to collect vehicle characteristic information of each target vehicle located in the merging area in a vehicle merging scenario; the merging area includes the main road and the ramp area that meets the preset distance condition from the main road.

[0060] Prediction module 12 is used to predict the merging intention of each target vehicle based on the road feature information of the vehicle merging scene and the feature information of each vehicle, and obtain the corresponding prediction results; the preset neural network is a neural network constructed based on a three-layer bidirectional long short-term memory network.

[0061] The planning module 13 is used to plan the driving trajectory of the current vehicle based on the prediction results, so as to obtain a target driving trajectory for avoiding the target vehicles when the vehicles merge into the scene.

[0062] Therefore, in this application, vehicles traveling on the main road can predict the intentions of each target vehicle located on the ramp to merge into the main road in advance in the vehicle merging scenario, and obtain the corresponding prediction results; then, by combining the prediction results, a driving trajectory that can avoid each target vehicle can be planned more accurately, thereby improving the driving safety of vehicles in the vehicle merging scenario.

[0063] In one specific embodiment, the device may further include:

[0064] The scene judgment module is used to determine whether the scene in which the current vehicle is driving matches the preset merged scene based on the surrounding map data of the current vehicle.

[0065] The merging area drawing module is used to draw a topology map based on the surrounding map data if the current vehicle driving scene matches the preset merging scene, so as to obtain the merging area topology map, and collect the vehicle feature information of each target vehicle located in the merging area based on the merging area topology map in the vehicle merging scene.

[0066] In one specific embodiment, the information acquisition module 11 may include:

[0067] The information acquisition unit is used to acquire attribute values ​​of each target vehicle located in the merging area through a camera and / or radar in a vehicle merging scenario, and to mark the attribute values ​​in the topology map of the merging area to obtain vehicle feature information of each target vehicle.

[0068] In one specific embodiment, the prediction module 12 may include:

[0069] The first feature extraction unit is used to perform feature extraction processing on the road feature information of the vehicle merging scene through the first bidirectional long short-term memory network to obtain the corresponding map tensor.

[0070] The second feature extraction unit is used to perform feature extraction processing on the feature information of each vehicle through the second bidirectional long short-term memory network to obtain the corresponding target tensor.

[0071] The intent prediction unit is used to perform input intent prediction on each of the target vehicles through a third bidirectional long short-term memory network and by utilizing the map tensor and the target tensor, and to obtain the corresponding prediction results.

[0072] In one specific embodiment, the planning module 13 may include:

[0073] An obstacle information labeling unit is used to label obstacle information at the location points corresponding to each target vehicle at each time point according to the prediction results, so as to plan the driving trajectory of the current vehicle based on the obstacle information.

[0074] In another specific embodiment, the planning module 13 may include:

[0075] The planning unit is used to plan the driving trajectory of the current vehicle based on the prediction results, the road feature information, and the vehicle feature information, so as to obtain a target driving trajectory for avoiding the target vehicles in the vehicle merging scenario.

[0076] Furthermore, embodiments of this application also disclose an electronic device, Figure 7 This is a structural diagram of an electronic device 20 according to an exemplary embodiment. The content of the diagram should not be construed as limiting the scope of this application.

[0077] Figure 7 This is a schematic diagram of the structure of an electronic device 20 provided in an embodiment of this application. Specifically, the electronic device 20 may include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. The memory 22 stores a computer program, which is loaded and executed by the processor 21 to implement the relevant steps in the vehicle trajectory planning method disclosed in any of the foregoing embodiments. Furthermore, the electronic device 20 in this embodiment may specifically be an electronic computer.

[0078] In this embodiment, the power supply 23 is used to provide operating voltage for each hardware device on the electronic device 20; the communication interface 24 can create a data transmission channel between the electronic device 20 and external devices, and the communication protocol it follows can be any communication protocol applicable to the technical solution of this application, and is not specifically limited here; the input / output interface 25 is used to acquire external input data or output data to the outside world, and its specific interface type can be selected according to specific application needs, and is not specifically limited here.

[0079] In addition, the memory 22, as a carrier for resource storage, can be a read-only memory, random access memory, disk or optical disk, etc. The resources stored thereon can include operating system 221, computer program 222, etc., and the storage method can be temporary storage or permanent storage.

[0080] The operating system 221 is used to manage and control the various hardware devices on the electronic device 20 and the computer program 222, which may be Windows Server, Netware, Unix, Linux, etc. In addition to including a computer program capable of performing the vehicle trajectory planning method executed by the electronic device 20 as disclosed in any of the foregoing embodiments, the computer program 222 may further include a computer program capable of performing other specific tasks.

[0081] Furthermore, this application also discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, it implements the aforementioned vehicle trajectory planning method. Specific steps of this method can be found in the corresponding content disclosed in the foregoing embodiments, and will not be repeated here.

[0082] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section.

[0083] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0084] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0085] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0086] The technical solutions provided in this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A vehicle trajectory planning method, characterized in that, Applicable to current vehicles traveling on the main road, including: In a vehicle merging scenario, vehicle characteristic information of each target vehicle located in the merging area is collected; the merging area includes the main road and the ramp area that meets the preset distance condition from the main road; Based on the road feature information of the vehicle merging scenario and the feature information of each vehicle, a preset neural network is used to predict the merging intention of each target vehicle and obtain the corresponding prediction results; the preset neural network is a neural network constructed based on a three-layer bidirectional long short-term memory network. Based on the prediction results, the current vehicle's driving trajectory is planned to obtain a target driving trajectory for avoiding each of the target vehicles in the vehicle merging scenario. The step of using a preset neural network to predict the merging intention of each target vehicle based on the road feature information of the vehicle merging scenario and the feature information of each vehicle, and obtaining corresponding prediction results, includes: The road feature information of the vehicle merging scene is processed by the first bidirectional long short-term memory network to obtain the corresponding map tensor. The feature information of each vehicle is extracted using a second bidirectional long short-term memory network to obtain the corresponding target tensor. By using a third bidirectional long short-term memory network and the map tensor and the target tensor, the merging intention of each target vehicle is predicted, and the corresponding prediction results are obtained.

2. The vehicle trajectory planning method according to claim 1, characterized in that, Before collecting vehicle feature information of each target vehicle located in the merging area in the vehicle merging scenario, the method further includes: Based on the surrounding map data of the current vehicle, determine whether the current driving scene of the vehicle matches the preset merging scene; If the current vehicle's driving scenario matches the preset merging scenario, a topology map is drawn based on the surrounding map data to obtain a merging area topology map, so as to collect vehicle feature information of each target vehicle located in the merging area based on the merging area topology map in the vehicle merging scenario.

3. The vehicle trajectory planning method according to claim 2, characterized in that, The process of collecting vehicle feature information of each target vehicle located in the merging area in a vehicle merging scenario includes: In a vehicle merging scenario, attribute values ​​of each target vehicle located in the merging area are collected by cameras and / or radar, and the attribute values ​​are marked on the topology map of the merging area to obtain vehicle feature information of each target vehicle.

4. The vehicle trajectory planning method according to any one of claims 1 to 3, characterized in that, The prediction results characterize the probability and location of the target vehicle merging from the ramp area into the main road at each time point.

5. The vehicle trajectory planning method according to claim 4, characterized in that, The step of planning the current vehicle's trajectory based on the prediction results includes: Based on the prediction results, obstacle information is marked at the location points corresponding to the target vehicles at each time point, so as to plan the current vehicle's driving trajectory based on the obstacle information.

6. The vehicle trajectory planning method according to claim 4, characterized in that, The step of planning the current vehicle's trajectory based on the prediction results to obtain a target trajectory for avoiding the target vehicles in the vehicle merging scenario includes: Based on the prediction results, road feature information, and vehicle feature information, the current vehicle's driving trajectory is planned to obtain a target driving trajectory for avoiding the target vehicles in the vehicle merging scenario.

7. A vehicle trajectory planning device, characterized in that, Applicable to current vehicles traveling on the main road, including: The information acquisition module is used to collect vehicle characteristic information of each target vehicle located in the merging area in a vehicle merging scenario; the merging area includes the main road and the ramp area that meets the preset distance condition from the main road; The prediction module is used to predict the merging intention of each target vehicle based on the road feature information of the vehicle merging scene and the feature information of each vehicle, and to obtain the corresponding prediction results; the preset neural network is a neural network constructed based on a three-layer bidirectional long short-term memory network. The planning module is used to plan the driving trajectory of the current vehicle based on the prediction results, so as to obtain a target driving trajectory to avoid the target vehicles when the vehicles merge into the scene. The prediction module includes: The first feature extraction unit is used to perform feature extraction processing on the road feature information of the vehicle merging scene through the first bidirectional long short-term memory network to obtain the corresponding map tensor. The second feature extraction unit is used to perform feature extraction processing on the feature information of each vehicle through the second bidirectional long short-term memory network to obtain the corresponding target tensor. The intent prediction unit is used to perform input intent prediction on each of the target vehicles through a third bidirectional long short-term memory network and by utilizing the map tensor and the target tensor, and to obtain the corresponding prediction results.

8. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor for executing the computer program to implement the vehicle trajectory planning method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, Used to store a computer program, which, when executed by a processor, implements the vehicle trajectory planning method as described in any one of claims 1 to 6.

Images