A path planning method and device

Abstract

Embodiments of the present application provide a path planning method and device, the method comprising: obtaining environment information of a first region in which a first vehicle is located, the first region being within a perception range of the first vehicle and comprising at least one lane, the first vehicle being located on a first lane of the at least one lane; determining, according to the environment information, lateral partition information of each lateral partition of a plurality of lateral partitions corresponding to each sampling position of at least one sampling position, each sampling position of the at least one sampling position being located in front of the first vehicle in a driving direction and being different in distance from the first vehicle, and a width of each lateral partition being less than a width of the first lane; and finally, planning a driving path of the first vehicle according to the environment information and the lateral partition information of each lateral partition of the plurality of lateral partitions corresponding to each sampling position. Since the width of the lateral partition is less than the width of the lane, the above method can improve the accuracy of the driving path of the first vehicle.

Description

Technical Field

[0001] This invention relates to the field of autonomous driving technology, and in particular to a path planning method and apparatus. Background Technology

[0002] In the field of autonomous driving technology, autonomous vehicles need to plan their own paths based on the surrounding roads and obstacles on the roads, and then drive along the planned paths.

[0003] Currently, existing technologies for path planning generally use lanes as the granularity, and the planned paths are usually the lanes that autonomous vehicles will travel in. Since lanes have a certain width, the accuracy of the paths planned by the above methods is relatively low. How to improve the accuracy of path planning is an urgent problem to be solved. Summary of the Invention

[0004] This application provides a path planning method and apparatus to improve the accuracy of path planning.

[0005] In a first aspect, embodiments of this application provide a path planning method, the method comprising: acquiring environmental information of a first area where a first vehicle is located, the first area being within the perception range of the first vehicle and including at least one lane, the first vehicle being located in the first lane of the at least one lane; then, based on the environmental information, determining the lateral partition information of each of a plurality of lateral partitions corresponding to each of at least one sampling location, each of the at least one sampling location being located ahead of the first vehicle's driving direction and at a different distance from the first vehicle's position, the width of each lateral partition being less than the width of the first lane; and finally, planning the driving path of the first vehicle based on the environmental information and the lateral partition information of each of the plurality of lateral partitions corresponding to each sampling location.

[0006] In this method, since the width of the lateral partition is smaller than the width of the lane, the driving path of the first vehicle planned based on environmental information and the lateral partition information of each of the multiple lateral partitions corresponding to each sampling position has high accuracy.

[0007] In one possible implementation, the lateral partition information includes a first offset distance of the lateral partition relative to the first vehicle in the lateral direction.

[0008] In one possible implementation, planning the driving path of the first vehicle based on environmental information and the lateral partition information of each of the multiple lateral partitions corresponding to each sampling location can include: determining the partition prediction result of each of the multiple lateral partitions corresponding to each sampling location based on the environmental information and the lateral partition information of each of the multiple lateral partitions corresponding to each sampling location. The partition prediction result includes a first prediction probability value and a first prediction deflection angle of each lateral partition relative to the driving direction of the first vehicle; and then planning the driving path of the first vehicle based on the partition prediction result of each of the multiple lateral partitions corresponding to each sampling location. This design provides a method for planning the driving path of the first vehicle, determining the lateral partition with a better lateral position among multiple lateral partitions based on the prediction probability value, and determining the angle at which the first vehicle passes through the lateral partition based on the deflection angle, making the planned path smoother.

[0009] In one possible implementation, planning the driving path of the first vehicle based on the partition prediction results of multiple lateral partitions corresponding to each sampling location may include: determining at least one target lateral partition corresponding to each sampling location based on the partition prediction results of multiple lateral partitions corresponding to each sampling location, and then planning the driving path of the first vehicle based on the lateral partition information of the at least one target lateral partition corresponding to each sampling location. This design provides a method for planning the driving path of a first vehicle.

[0010] In one possible implementation, determining at least one target horizontal partition corresponding to each sampling location based on the partition prediction results of each of the multiple horizontal partitions corresponding to each sampling location may include: selecting the horizontal partition with the highest first prediction probability value among the multiple horizontal partitions corresponding to the first sampling location as the first target horizontal partition corresponding to the first sampling location. In another possible implementation, at least one horizontal partition that satisfies a preset partitioning relationship with the first target horizontal partition may also be selected as the second target horizontal partition. This design provides multiple methods for determining target horizontal partitions. By determining the target horizontal partitions, path planning for the first vehicle can be performed based on the partition prediction results of the target horizontal partitions, resulting in a more accurate planned path.

[0011] In one possible implementation, planning the driving path of the first vehicle based on the lateral partition information of at least one target lateral partition corresponding to each sampling position may include: determining the position information and deflection angle information of the prediction point corresponding to each sampling position based on the lateral partition information of at least one target lateral partition corresponding to each sampling position, and then planning the driving path of the first vehicle based on the position information and deflection angle information of the prediction point corresponding to each sampling position. This design provides a method for planning the driving path of a first vehicle.

[0012] In one possible implementation, environmental information includes road topology information and traffic participant information.

[0013] In one possible implementation, the multiple lateral partitions corresponding to each sampling location are obtained by segmenting at least one lane at each sampling location in the lateral direction relative to the first vehicle.

[0014] Secondly, embodiments of this application provide a path planning device that implements the method example described in the first aspect. This function can be implemented in hardware or by hardware executing corresponding software. The hardware or software includes one or more modules corresponding to the aforementioned function. In one possible design, the path planning device includes an acquisition module and a processing module. The acquisition module is used to acquire environmental information of a first area where a first vehicle is located. The first area is within the perception range of the first vehicle, and the first area includes at least one lane, with the first vehicle located in the first lane of the at least one lane. The processing module is used to determine, based on the acquired environmental information, the lateral partition information of each of the plurality of lateral partitions corresponding to each of the at least one sampling position. Each of the at least one sampling position is located ahead of the first vehicle's driving direction and at a different distance from the first vehicle's position. The width of each lateral partition is less than the width of the first lane. The processing module is also used to plan the driving path of the first vehicle based on the environmental information and the lateral partition information of each of the plurality of lateral partitions corresponding to each sampling position.

[0015] The path planning device can perform the corresponding functions of the method example in the first aspect described above. The beneficial effects can be found in the description of the first aspect, and will not be repeated here.

[0016] Thirdly, this application provides a path planning apparatus, which includes a processor and a memory, wherein the memory is used to store computer program instructions. Optionally, the processor is coupled to the memory. When the processor executes the computer program instructions, it causes the path planning apparatus to perform the methods described in the first aspect or any possible implementation thereof. In one example, the path planning apparatus further includes a transceiver, which may be a transceiver device integrated with data transmission and reception functions. In another example, the transceiver may also consist of a transmitter and a receiver, wherein the transmitter is used to transmit data and the receiver is used to receive data.

[0017] Fourthly, this application provides a possible route planning system, which includes the route planning device of the second or third aspect. Exemplarily, the number of route planning devices included in the route planning system can be one or more, and this application does not limit this. The route planning system can be deployed on a vehicle.

[0018] Fifthly, this application also provides a vehicle that includes the above-described path planning device or the above-described path planning system.

[0019] Sixthly, this application provides a computer program product including computer program instructions that, when executed on a computer, cause the computer to perform the method described in the first aspect or any possible implementation thereof.

[0020] In a seventh aspect, this application provides a computer-readable storage medium storing computer program instructions that, when executed by a computer, cause the computer to perform the method described in the first aspect or any possible implementation thereof.

[0021] Eighthly, this application also provides a chip coupled to a memory for reading computer program instructions stored in the memory and executing the method in any possible implementation of the first aspect above.

[0022] Ninthly, this application also provides a chip system including a processor for supporting a computer device in implementing the methods of the first aspect or any possible implementation thereof. In one possible implementation, the chip system further includes a memory for storing program instructions and data necessary for the computer device. The chip system may be composed of chips or may include chips and other discrete devices.

[0023] These or other aspects of this application will become more apparent in the following description of the embodiments. Attached Figure Description

[0024] Figure 1 A schematic diagram of a system structure provided in an embodiment of this application;

[0025] Figure 2 A flowchart illustrating a method for training a partition prediction model provided in an embodiment of this application;

[0026] Figure 3 A flowchart illustrating a method for obtaining first sample data provided in an embodiment of this application;

[0027] Figure 4 This is a schematic diagram of the structure of a sample region provided in an embodiment of this application;

[0028] Figure 5 This is a schematic diagram of the structure of a sample region provided in an embodiment of this application;

[0029] Figure 6 This is a schematic diagram of a lane segmentation structure provided in an embodiment of this application;

[0030] Figure 7a This is a schematic diagram of a lane segmentation structure provided in an embodiment of this application;

[0031] Figure 7b This is a schematic diagram of a lane segmentation structure provided in an embodiment of this application;

[0032] Figure 8 This is a schematic diagram of the structure of a sample region provided in an embodiment of this application;

[0033] Figure 9 This is a schematic diagram of a first sampling position provided in an embodiment of this application;

[0034] Figure 10 A flowchart illustrating a method for determining horizontal partition information provided in an embodiment of this application;

[0035] Figure 11a This is a schematic diagram of a segmentation point provided in an embodiment of this application;

[0036] Figure 11b This is a schematic diagram of a segmentation point provided in an embodiment of this application;

[0037] Figure 12 This is a schematic diagram of the structure of a first deflection angle label provided in an embodiment of this application;

[0038] Figure 13 This is a schematic diagram of the structure of a partitioning prediction model provided in an embodiment of this application;

[0039] Figure 14This is a schematic diagram of the structure of a partitioning prediction model provided in an embodiment of this application;

[0040] Figure 15 This application provides a schematic diagram of the structure of a partitioning prediction model and a loss function calculation model.

[0041] Figure 16 A flowchart illustrating a path planning method provided in an embodiment of this application;

[0042] Figure 17 A flowchart illustrating a method for planning a driving route provided in an embodiment of this application;

[0043] Figure 18 A flowchart illustrating a method for planning a driving route provided in an embodiment of this application;

[0044] Figure 19 A flowchart illustrating a method for determining the location information of a predicted point, provided in an embodiment of this application;

[0045] Figure 20 A schematic diagram of the structure of a prediction point provided in an embodiment of this application;

[0046] Figure 21 An exemplary schematic diagram of a possible path planning device provided in an embodiment of this application is shown.

[0047] Figure 22 An exemplary schematic diagram of a possible path planning device provided in an embodiment of this application is shown. Detailed Implementation

[0048] The technical solutions in the embodiments of this application will now be described with reference to the accompanying drawings.

[0049] Figure 1 This is a schematic diagram of an autonomous driving system that can be applied to an embodiment of this application, including an environmental perception module and a path planning module, wherein the path planning module is the core module for path planning in the embodiment of this application.

[0050] Environmental perception module: This module is responsible for sensing the surrounding environment and obtaining the current driving environment status of the vehicle. Specifically, it integrates information transmitted from various sensors to determine the vehicle's status, including its position, speed, and yaw angle; it also determines the surrounding environment status, including the position of static targets (lane boundary lines, etc.) and the position and speed of dynamic targets (other vehicles and pedestrians, etc.), and transmits this information to the path planning module.

[0051] Path planning module: Based on the vehicle's status and the status of other vehicles, it plans a path that the vehicle can travel.

[0052] The driving decision-making scheme provided in this application embodiment can be applied to applications such as Figure 1 The illustrated autonomous driving system includes an onboard terminal, onboard computing platform, domain controller, and autonomous vehicle. In the subsequent description of the embodiments in this application, applications such as... Figure 1 The autonomous driving system shown is illustrated using an autonomous vehicle as an example.

[0053] In this embodiment of the application, a method for training a partitioning prediction model is provided, wherein the first vehicle is an autonomous vehicle equipped with an autonomous driving system. See also Figure 2 The specific training process of the partition prediction model is as follows:

[0054] Step 201: The first vehicle acquires at least one set of first sample data.

[0055] In the embodiments of this application, for any set of first sample data, it can be done by, as follows: Figure 3 The acquisition method flow shown includes the following steps.

[0056] Step 301: The first vehicle acquires environmental information of the sample area where it is located. This can be understood as the first vehicle acquiring environmental information of the sample area where it is located at time i, where time i can be any time before the current time.

[0057] In this embodiment, the sample region is within the perception range of the first vehicle. The sample region is determined based on the sample position information corresponding to the location of the first vehicle at time i, and at least one second preset distance. Two possible implementations for determining the sample region are described below.

[0058] In one possible implementation, the sample region is defined by using the sample position information corresponding to the location of the first vehicle at time i as the center coordinates of a circle, and a second preset distance as the radius. The sample region determined by the above method includes at least one lane, and the first vehicle is located in the first lane of the at least one lane.

[0059] like Figure 4The diagram shows lane 1 and lane 2, both of which are north-south straight lanes. Lane 1 is defined by a hard lane boundary line 1 and a soft lane boundary line, while lane 2 is defined by a soft lane boundary line and a hard lane boundary line 2. Both hard lane boundary lines 1 and 2 are boundary lines that vehicles cannot cross, while the soft lane boundary line is a boundary line that vehicles can cross. A first vehicle is located in north-south straight lane 1. A sample area is defined using the sample position information corresponding to the first vehicle's position at time i as the center and a second preset distance D as the radius. This sample area includes lanes 1 and 2, with lane 1 being the first lane.

[0060] The second possible implementation uses the sample position information corresponding to the location of the first vehicle at time i as the coordinates of the center point of the rectangle, a first preset width as the width of the rectangle, and a first preset length as the length of the rectangle. The determined rectangular area is the sample area. Here, the first preset width and the first preset length are both second preset distances, and the first preset width and the first preset length can be the same or different. The sample area determined by the above method includes at least one lane, and the first vehicle is located in the first lane of the at least one lane.

[0061] like Figure 5 As shown, the first vehicle is located in lane 1, which is a straight north-south direction. The coordinates of the center point of the rectangle are taken as the sample position information corresponding to the position of the first vehicle at time i. The first preset width W1 is the width of the rectangle, and the first preset length L1 is the length of the rectangle. Based on the above, the sample area can be determined. The sample area includes lane 1 and lane 2. The lane 1 where the first vehicle is located is the first lane.

[0062] In this embodiment of the application, a map application software is installed in the first vehicle. The map application software can open a high-precision map. After determining the sample area, the environmental information of the sample area can be obtained from the high-precision map. The environmental information includes road topology information and traffic participant information.

[0063] Road topology information includes the road topology structure of the sample area, such as the number of lanes, intersections, changes in the number of lanes (decreased number of lanes, increased number of lanes, etc.), lane direction (such as one-way streets, two-way streets, etc.), lane boundaries, lane lines (double yellow lines, dashed and solid lines, etc.), speed limit information, lane type (north-south straight lanes, east-west straight lanes, left-turn lanes, right-turn lanes, etc.).

[0064] In one possible implementation, road topology information can be determined through a variety of possible implementations.

[0065] With the location of the first vehicle at time i as the origin O(0,0), the direction of travel of the first vehicle at time i as the y-axis, and the direction perpendicular to the direction of travel of the first vehicle at time i as the x-axis, a rectangular coordinate system xOy is constructed. This rectangular coordinate system xOy is the vehicle body coordinate system corresponding to the first vehicle at time i. The vehicle body coordinate system corresponding to the first vehicle at time i can also be determined in other ways, which are not limited here. For ease of description later, the vehicle body coordinate system corresponding to the first vehicle at time i can be simply referred to as the first coordinate system.

[0066] First, each lane in the sample area is longitudinally segmented according to a first preset distance to obtain at least one lane segment. In one possible implementation, when segmenting the lanes in the sample area, segmentation can be performed along the lanes according to the first preset distance to obtain at least one lane segment. For example, such as... Figure 6 As shown, the first preset distance is 5 meters. The sample area includes lane 1. Along lane 1 in the sample area, it is divided into 3 lane segments according to 5 meters.

[0067] The coordinates of each lane segment in at least one lane within the sample area are determined in a first coordinate system. Generally, the coordinates of any point within a lane segment in the first coordinate system can be used as the coordinates of the lane segment in the first coordinate system. In this embodiment, the coordinates of the center point of the lane segment in the first coordinate system can be used as the coordinates of the lane segment in the first coordinate system.

[0068] Determine the first deflection angle of each lane segment in at least one lane in the sample area relative to the travel direction of the first vehicle at time i. The value of the first deflection angle is in the range of -180°. ○ ~180 ○ (where -180) ○ and 180 ○ (The first deflection angle is the same). In one possible implementation, the first deflection angle can be determined by the following possible implementation: taking the angle between the lane-defined driving direction corresponding to each lane segment and the driving direction of the first vehicle at time i as the first deflection angle. If the lane-defined driving direction corresponding to each lane segment deflects to the right relative to the driving direction of the first vehicle at time i, then the first deflection angle is at 0. ○ ~180 ○ Between; if the lane-defined driving direction corresponding to each lane segment deviates to the left relative to the driving direction of the first vehicle at time i, then the first deflection angle is located at -180. ○ ~0 ○ between.

[0069] like Figure 7aThe lane segment in lane 1 shown has a lane-defined driving direction that is the same as the driving direction of the first vehicle at time i. The angle between the lane-defined driving direction of the lane segment and the driving direction of the first vehicle at time i is 0, that is, the first deflection angle is 0.

[0070] like Figure 7b The lane segment in lane 1 shown has an angle d1 between the lane's designated driving direction and the driving direction of the first vehicle at time i, which is the first deflection angle d1.

[0071] Determine the speed limit information for each lane segment in at least one lane within the sample area. For example, the speed limit information for a lane segment is 60 km / h.

[0072] Determine the lane type for each lane segment in at least one lane within the sample area. Lane types include north-south straight lanes, east-west straight lanes, left-turn lanes, right-turn lanes, etc., without limitation.

[0073] In this embodiment of the application, the coordinates of the lane segment in the sample area in the first coordinate system, the first deflection angle of the lane segment relative to the driving direction of the first vehicle at time i, the speed limit information of the lane segment, and the lane type of the lane segment are all used as lane features. The number of lane features in this application is 4.

[0074] If the number of lanes in the sample area is M, the number of lane segments in each lane is P, and the number of lane features is F, then the road topology information of the sample area where the first vehicle is located at time i can be an M*P*F matrix, where the values ​​in the matrix are the values ​​of the lane features corresponding to each lane segment in the sample area, and M, P, and F are all integers greater than 0.

[0075] It should be understood that when dividing a lane, if the number of lane segments obtained is less than P, a preset value of -1 can be used as the lane feature.

[0076] In this embodiment of the application, traffic participants include a first vehicle in the sample area and obstacles. The obstacles include static targets in the sample area, such as road debris (e.g., stones), water barriers, cones, etc., and dynamic targets in the sample area, such as social vehicles and pedestrians. The social vehicles include motor vehicles and non-motor vehicles (e.g., bicycles, electric vehicles, motorcycles, etc.).

[0077] In this embodiment of the application, a sensor is installed in the first vehicle, which can obtain speed and location information of traffic participants in the lane.

[0078] Traffic participant information includes information about the first vehicle, such as its position, speed, acceleration, steering angle, and historical driving trajectory. The steering angle can be the angle of deflection of the first vehicle's driving direction relative to the lane line of the first lane it is in.

[0079] Traffic participant information also includes obstacle information, including the type, location, speed, acceleration, steering angle, and historical driving trajectory of the obstacle. The steering angle can be the deflection angle of the obstacle relative to the direction of travel of the first vehicle or the deflection angle relative to the lane line of the lane in which it is located.

[0080] Traffic participant information may also include the speed information of traffic participants within a third historical time period. Generally, multiple third historical moments are obtained from the third historical time period, and the speed information of each traffic participant at each of the multiple third historical moments is determined. The multiple third historical moments and moment i satisfy a first time preset relationship. For example, the first time preset relationship can be understood as the multiple third historical moments differing from each other by a first preset time period, and the largest third historical moment among the multiple third historical moments differing from moment i by a first preset time period.

[0081] In one possible implementation, traffic participant information can be determined through a variety of possible methods.

[0082] Determine the coordinates of at least one obstacle in the sample area in the first coordinate system. The first coordinate system has been described above and will not be repeated here.

[0083] Determine the second deflection angle of at least one obstacle in the sample area relative to the travel direction of the first vehicle at time i, wherein the value of the second deflection angle ranges from -180°. ○ ~180 ○ (where -180) ○ and 180 ○ (The second deflection angle is the same). In one possible implementation, the second deflection angle can be determined as follows: If the obstacle is a vehicle or pedestrian, the angle between the obstacle's direction of travel at time i and the first vehicle's direction of travel at time i is taken as the second deflection angle. If the obstacle is road debris, since road debris has no direction of travel, the second deflection angle is 0. ○ .like Figure 8 In the sample area shown, when the obstacle is a motor vehicle, the obstacle's direction of travel at time i is the same as the first vehicle's direction of travel at time i. Therefore, the angle between the obstacle's direction of travel at time i and the first vehicle's direction of travel at time i is 0. ○ That is, the second deflection angle is 0. ○ .

[0084] Determine the velocity information of at least one obstacle in the sample area at time i. If the obstacle is a vehicle or pedestrian, its velocity at time i is non-zero. If the obstacle is road debris, its velocity at time i is zero.

[0085] Determine the speed information of at least one obstacle in the sample area within a third historical time period. Generally, multiple third historical moments are obtained from the third historical time period, and the speed information of each obstacle at each of the multiple third historical moments is determined. Here, the third historical time period is the time period before moment i, the third historical moment is the moment before moment i, and the multiple third historical moments and moment i satisfy a first time preset relationship. In one possible implementation, the first time preset relationship can be understood as the multiple third historical moments differing from each other by a first preset time period, and the largest third historical moment differing from moment i by a first preset time period. For example, three third historical moments are obtained from the third historical time period: third historical moment 1, third historical moment 2, and third historical moment 3. Third historical moment 1 is less than third historical moment 2, third historical moment 2 is less than third historical moment 3, third historical moment 1 differs from third historical moment 2 by 0.01 seconds, third historical moment 2 differs from third historical moment 3 by 0.01 seconds, and third historical moment 3 differs from moment i by 0.01 seconds.

[0086] Determine the obstacle type of each obstacle in at least one obstacle in the sample area, wherein the obstacle type includes motor vehicles, non-motor vehicles, pedestrians, road debris, etc., without limitation.

[0087] Determine the coordinates of at least one obstacle in the sample area within a third historical time period, respectively, in a first coordinate system. Generally, multiple third historical moments are obtained from the third historical time period, and the coordinates of each obstacle's position in the first coordinate system are determined at each of the multiple third historical moments.

[0088] For any third historical moment among multiple third historical moments, and any obstacle among at least one obstacle, if the position of the obstacle cannot be determined at that third historical moment, the coordinates of the obstacle's position at that third historical moment in the first coordinate system can be set to preset coordinates, such as (-1, -1).

[0089] Determine the speed information of the first vehicle at time i.

[0090] The vehicle type of the first vehicle is determined. The vehicle type of the first vehicle may include motor vehicles, non-motor vehicles, etc., and is not limited here.

[0091] Determine the position of the first vehicle within the third historical time period using its coordinates in the first coordinate system. Generally, multiple third historical moments are obtained from the third historical time period, and the position of the first vehicle at each of these multiple third historical moments is determined using its coordinates in the first coordinate system.

[0092] Determine the speed information of the first vehicle within the third historical time period. Generally, multiple third historical moments are obtained from the third historical time period, and the speed information of the first vehicle at each of the multiple third historical moments is determined.

[0093] In this embodiment, the following are considered as traffic participant features: the coordinates of the obstacle in the first coordinate system, the second deflection angle of the obstacle relative to the driving direction of the first vehicle at time i, the speed information of the obstacle at time i, the obstacle type, the speed information of the obstacle in the third historical time period, the coordinates of the obstacle's position in the first coordinate system in the third historical time period, the speed information of the first vehicle at time i, the type of the first vehicle, the coordinates of the first vehicle's position in the first coordinate system in the third historical time period, and the speed information of the first vehicle in the third historical time period. The number of traffic participant features in this application is 10.

[0094] If the number of traffic participants in the sample area is N, the number of reference times (including multiple third historical times and time i) is T, and the number of traffic participant features is H, then the traffic participant information in the sample area can be an N*T*H matrix, where the values ​​in the matrix are the traffic participant feature values ​​corresponding to each traffic participant at each reference time, and N, T, and H are all integers greater than 0.

[0095] Step 302: The first vehicle determines the lateral partition information of each lateral partition in the multiple lateral partitions corresponding to each sampling location in at least one sampling location, based on the environmental information.

[0096] In the embodiments of this application, each of the at least one sampling position is located in front of the first vehicle in the direction of travel, and the at least one sampling position is at a different distance from the first vehicle at time i.

[0097] For the first sampling position, which is any one of at least one sampling position, the first sampling position can be determined by the following possible implementation: taking the position at a third preset distance from the position of the first vehicle at time i as the first sampling position.

[0098] In this embodiment of the application, different sampling positions can be determined by a third preset distance that is different from the position of the first vehicle at time i.

[0099] It should be understood that the third preset distance can be a preset value, such as 100 meters; the third preset distance can be determined based on the speed information of the first vehicle at time i and the second preset time period, that is, the speed of the first vehicle at time i multiplied by the second preset time period is the third preset distance. For example, if the speed of the first vehicle at time i is 40 km / h and the second preset time period is 3 seconds, then the third preset distance is 40 km / h * 3 seconds = (100 / 3) meters.

[0100] In this embodiment of the application, the first sampling position is defined as the location at a third preset distance from the position of the first vehicle at time i. This can include the following possible implementation methods: such as... Figure 9 As shown, firstly, a third reference line is determined, which includes the position of the first vehicle at time i and is parallel to the center line of the first lane. Then, along the third reference line, a location 300 meters away from the position of the first vehicle at time i is taken as the first sampling position.

[0101] In this embodiment of the application, the third reference line may also be other reference lines including the position of the first vehicle at time i, which is not limited here.

[0102] In this embodiment of the application, after the sampling location is determined, multiple lateral partitions corresponding to each sampling location are determined. The multiple lateral partitions corresponding to each sampling location are obtained by segmenting at least one lane of each sampling location in the lateral direction relative to the first vehicle. Each sampling location is within the multiple lateral partitions corresponding to each sampling location.

[0103] Determine multiple horizontal partitions corresponding to each sampling location and the horizontal partition information, which may include, for example: Figure 10 The following steps are shown:

[0104] Step 1001: The first vehicle determines a first reference line corresponding to at least one sampling position, wherein the first reference line is perpendicular to the lane center line of the first lane.

[0105] In this embodiment, a second reference line is first determined. This second reference line can be the lane centerline of the first lane, or it can be a line parallel to the lane centerline of the first lane, such as a third reference line. Then, at least one first reference line is determined that is perpendicular to the second reference line and contains at least one sampling position, with each sampling position corresponding to one first reference line.

[0106] The second reference line is a line parallel to the center line of the first lane. This can be understood as the angle between the second reference line and the center line of the first lane being less than a preset angle.

[0107] Step 1002: The first vehicle segments the first reference line corresponding to each sampling position to obtain multiple segmentation points, and takes the area between two adjacent segmentation points as a horizontal partition, resulting in multiple horizontal partitions corresponding to each sampling position. The width of each horizontal partition is less than the width of the first lane. The width of each horizontal partition is a second preset width W2.

[0108] In this embodiment of the application, for the first sampling position, which is any one of at least one sampling position, determining the segment point corresponding to the first sampling position may include the following possible implementation methods:

[0109] In a first possible implementation, on the first reference line, the point at a distance of n*W2 from the first sampling position is a segmentation point. The first sampling position can also be used as a segmentation point, where n is an integer greater than 0. For example... Figure 11a As shown, the first sampling position is point O. On the first reference line, points a1 and a2 are at a distance W2 from point O, points b1 and b2 are at a distance 2*W2 from point O, and points c1 and c2 are at a distance 3*W2 from point O. Among them, points O, a1, a2, b1, b2, c1, and c2 are all segmentation points.

[0110] The second possible implementation involves first determining a point on the first reference line that is (W2) / 2 distances from the first sampling position as the starting point. Then, on the first reference line, points on the first reference line that are n*W2 distances from the starting point are designated as segmentation points. The starting point can also be used as a segmentation point, where n is a positive integer. For example... Figure 11b As shown, the first sampling position is point O. Points a1 and a2 are defined on the first reference line that are (W2) / 2 away from the first sampling position. Point b1 is located W2 away from point a1, point c1 is located 2*W2 away from point a1, point b2 is located W2 away from point a2, and point c2 is located 2*W2 away from point a2. Points a1, a2, b1, b2, c1, and c2 are all segmentation points.

[0111] The width between the segment points determined by the above methods is the same. Other methods can also be used to determine the width between the segment points, which are not specifically limited here.

[0112] Step 1003: The first vehicle acquires the lateral partition information of multiple lateral partitions corresponding to each sampling position in at least one sampling position.

[0113] In this embodiment of the application, the horizontal partitioning information can be determined in the following ways:

[0114] Determine the width information of multiple horizontal partitions corresponding to each sampling position.

[0115] Determine the lane type of the lane in which the multiple lateral partitions corresponding to each sampling location are located. If the lateral partition includes a portion of two lanes, then select the lane type of any lane as the lane type of the lane in that lateral partition.

[0116] Determine the first offset distance (i.e., the first offset distance relative to the third reference line) of each of the multiple lateral partitions corresponding to each sampling position relative to the first vehicle in the lateral direction. Generally, the vertical distance of the center point of the lateral partition relative to the third reference line can be used as the first offset distance.

[0117] Determine the second offset distance of each of the multiple lateral partitions corresponding to each sampling location relative to the lane centerline of the first lane. Generally, the vertical distance of the center point of the lateral partition relative to the lane centerline of the first lane can be used as the second offset distance.

[0118] For each sampling location, determine the third deflection angle of multiple lateral partitions relative to the first vehicle's direction of travel at time i. The value range of the third deflection angle is -180°. ○ ~180 ○ (where -180) ○ and 180 ○ (The third deflection angle is the same). The third deflection angle can be determined by the following possible implementation: the angle between the lane-defined driving direction of the lane corresponding to each lateral section and the driving direction of the first vehicle at time i is taken as the third deflection angle. If the lane-defined driving direction of the lane corresponding to each lateral section deflects to the right relative to the driving direction of the first vehicle at time i, then the third deflection angle is 0. ○ ~180 ○ Between; if the lane-defined driving direction of each lane corresponding to each lateral partition is deflected to the left relative to the driving direction of the first vehicle at time i, then the third deflection angle is -180°. ○ ~0 ○ between.

[0119] In this embodiment, the lane type of the lane where the lateral partition is located, the second offset distance of the lateral partition relative to the lane centerline of the first lane, the first offset distance of the lateral partition relative to the third reference line, and the third deflection angle of the lateral partition relative to the driving direction of the first vehicle at time i are all used as partition features. The number of partition features in this application is 4.

[0120] If the number of horizontal partitions corresponding to at least one sampling position is K, and the number of partition features is G, then the horizontal partition information can be a K*G matrix, where the values ​​in the matrix are the values ​​of the partition features corresponding to each horizontal partition, and K and G are both integers greater than 0.

[0121] In this embodiment of the application, each group of first sample data has a corresponding first label. For any group of first sample data, the first label corresponding to the first sample data includes at least one first partition label and a first deflection angle label corresponding to each sampling position.

[0122] For the first sampling position (which is any one of at least one sampling position), the intersection of the actual driving trajectory of the first vehicle and the first reference line is determined and denoted as the first reference point. The number of the lateral partition corresponding to the first sampling position where the first reference point falls is the first partition label. The deflection angle of the first vehicle's driving direction at the first reference point relative to the first vehicle's driving direction at time i is the first deflection angle label, and the value range of the first deflection angle label is -180°. ○ ~180 ○ (where -180) ○ and 180 ○ (The label for the first deflection angle is the same).

[0123] The explanation will be based on the example of multiple horizontal partitions corresponding to the first sampling position determined in the first possible implementation, such as... Figure 12 As shown, the horizontal partition between segment point o and segment point a1 is labeled as 1, the horizontal partition between segment point a1 and segment point b1 is labeled as 2, the horizontal partition between segment point b1 and segment point c1 is labeled as 3, the horizontal partition between segment point o and segment point a2 is labeled as -1, the horizontal partition between segment point a2 and segment point b2 is labeled as -2, and the horizontal partition between segment point b2 and segment point c2 is labeled as -3. The intersection of the actual driving trajectory of the first vehicle and the first reference line is the first reference point. The first reference point falls in the horizontal partition numbered 1; therefore, the first partition is labeled as 1. The deflection angle of the first vehicle's driving direction at the first reference point relative to the first vehicle's driving direction at time i is d2; therefore, the first deflection angle is labeled as d2.

[0124] Step 202: The first vehicle inputs at least one set of first sample data into the partition prediction model to obtain the first prediction result of each of the multiple horizontal partitions corresponding to at least one sampling position of each set of first sample data.

[0125] In the embodiments of this application, such as Figure 13 The partition prediction model shown includes a feature extraction module, a classification module, and a regression module. The feature extraction module includes a first fusion feature extraction module, a second fusion feature extraction module, and a third fusion feature extraction module.

[0126] Traffic participant information and road topology information from at least one set of first sample data are input into a first fusion feature extraction module to obtain a first fusion feature vector and a second fusion feature vector, respectively. The first fusion feature vector and the lateral partition information of each lateral partition corresponding to each sampling position in at least one set of first sample data are input into a second fusion feature extraction module to obtain a third fusion feature vector. The second and third fusion feature vectors are input into a third fusion feature extraction module to obtain a target fusion feature vector. The target fusion feature vector is input into a classification module and a regression module, respectively, to obtain a first prediction result for each lateral partition corresponding to at least one sampling position in each set of first sample data. Taking the first prediction result of each lateral partition corresponding to a sampling position in one set of first sample data as an example, the first prediction result of each lateral partition corresponding to a sampling position in this set of first sample data may include a first prediction probability value of each lateral partition corresponding to the sampling position, and a first prediction deflection angle of each lateral partition relative to the driving direction of the first vehicle at time i. The classification module outputs the first prediction probability value of each lateral partition, and the regression module outputs the first prediction deflection angle of each lateral partition relative to the driving direction of the first vehicle at time i.

[0127] It should be understood that the first fusion feature extraction module, the second fusion feature extraction module, and the third fusion feature extraction module can be cross-attention mechanisms, and no restrictions are imposed here.

[0128] like Figure 14 The partition prediction model shown may further include a first graph neural network and a second graph neural network. First, traffic participant information from at least one set of first sample data is input into the first graph neural network to obtain a first feature vector. Road topology information from at least one set of first sample data is input into the second graph neural network to obtain a second feature vector. The first and second feature vectors are then input into a first fusion feature extraction module to obtain a first fusion feature vector and a second fusion feature vector, respectively. The first fusion feature vector and the lateral partition information of each of the multiple lateral partitions corresponding to each sampling position in at least one set of first sample data are input into the second fusion feature extraction module to obtain a third fusion feature vector. The second and third fusion feature vectors are then input into the third fusion feature extraction module to obtain a target fusion feature vector. The target fusion feature vector is then input into a classification module and a regression module to obtain a first prediction result for each of the multiple lateral partitions corresponding to at least one sampling position in each set of first sample data.

[0129] It should be understood that the first and second graph neural networks can be graph neural networks (GNNs), and this is not a limitation.

[0130] Step 203: The first vehicle inputs the first prediction result and the first label of at least one set of first sample data into the loss function calculation model to obtain the loss value between the first prediction result and the first label of each set of first sample data. The parameters in the partition prediction model are updated according to the loss value between the first prediction result and the first label of each set of first sample data until the training is completed, and the trained partition prediction model is obtained.

[0131] like Figure 15 The model structure shown includes a partition prediction model and a loss function calculation model. The loss function calculation model includes a cross-entropy loss function module and a mean square error (MSE) loss function module.

[0132] For each of the first sample data in at least one set of first sample data, the following calculation is performed at the first sampling position.

[0133] The cross-entropy loss function module receives the first predicted probability value of each horizontal partition corresponding to the first sampling position from the classification module, as well as the first partition label corresponding to the first sampling position. It then uses the cross-entropy loss function to calculate the loss value between the first predicted probability value and the first partition label for each horizontal partition corresponding to the first sampling position of the first sample data, and updates the parameters in the partition prediction model based on this loss value.

[0134] The mean squared error loss function module receives the first predicted deflection angle of each lateral partition corresponding to the first sampling position relative to the driving direction of the first vehicle at time i, as well as the first deflection angle label corresponding to the first sampling position, output by the regression module. Using the mean squared error loss function, it calculates the loss value of the first predicted deflection angle and the first deflection angle label of each lateral partition corresponding to the first sampling position relative to the driving direction of the first vehicle at time i, and updates the parameters in the partition prediction model based on this loss value.

[0135] In this embodiment of the application, the parameters in the partition prediction model are updated based on the first predicted probability value and the loss value of the first partition label of each horizontal partition corresponding to at least one sampling position, the first predicted deflection angle of each horizontal partition corresponding to at least one sampling position relative to the driving direction of the first vehicle at time i and the loss value of the first deflection angle label, which can train the partition prediction model faster.

[0136] This application provides a flowchart of a path planning method, such as... Figure 16 As shown, the following steps may be included:

[0137] Step 1601: The first vehicle obtains environmental information of the first area where the first vehicle is located. The first area is within the perception range of the first vehicle and includes at least one lane. The first vehicle is located in the first lane of the at least one lane.

[0138] In this embodiment of the application, the first region is within the perception range of the first vehicle. The following examples illustrate possible implementations for determining the first region.

[0139] For example, the area within a circle defined by the coordinates of the current location of the first vehicle as the center and the second preset distance as the radius is the first region.

[0140] For example, using the current location of the first vehicle as the coordinates of the center point of the rectangle, the first preset width as the width of the rectangle, and the first preset length as the length of the rectangle, the defined rectangular area is called the first area. Here, the first preset width and the first preset length are both second preset distances, and the first preset width and the first preset height can be the same or different.

[0141] For example, the first area is the area that the first vehicle can perceive through onboard sensors or onboard sensing systems, or it can be the area perceived based on vehicle to everything (V2X), or it can be the area that can be perceived by a combination of both.

[0142] In this embodiment of the application, after determining the first region, environmental information of the first region can be obtained, wherein the environmental information includes road topology information and traffic participant information.

[0143] Road topology information includes the road topology structure of the first area, such as the number of lanes, intersections, changes in the number of lanes (decrease in the number of lanes, increase in the number of lanes, etc.), lane direction (such as one-way streets, two-way streets, etc.), lane boundaries, lane lines (double yellow lines, dashed and solid lines, etc.), speed limit information, lane type (north-south straight lanes, east-west straight lanes, left-turn lanes, right-turn lanes, etc.).

[0144] Traffic participant information includes information about the first vehicle, such as its position, speed, acceleration, steering angle, and historical driving trajectory. The steering angle can be the angle of deflection of the first vehicle's driving direction relative to the lane line of the first lane it is in.

[0145] Traffic participant information also includes obstacle information, including the type, location, speed, acceleration, steering angle, and historical driving trajectory of the obstacle. The steering angle can be the deflection angle of the obstacle relative to the direction of travel of the first vehicle or the deflection angle relative to the lane line of the lane in which it is located.

[0146] Traffic participant information may also include the speed information of traffic participants within a first historical time period. Generally, multiple first historical moments are obtained from the first historical time period, and the speed information of each traffic participant at each of the multiple first historical moments is determined. The multiple first historical moments and the current moment satisfy a first time preset relationship. For example, the first time preset relationship can be understood as the multiple first historical moments differing from each other by a first preset time period, and the largest first historical moment among the multiple first historical moments differing from the current moment by a first preset time period.

[0147] Step 1602: The first vehicle determines the lateral partition information of each lateral partition in the multiple lateral partitions corresponding to each sampling location in at least one sampling location, based on the environmental information.

[0148] In the embodiments of this application, each of the at least one sampling position is located in front of the first vehicle in the direction of travel, and each of the at least one sampling position is at a different distance from the location of the first vehicle at the current moment.

[0149] Regarding the first sampling position, which can be any one of at least one sampling position, the first sampling position can be determined through the following possible implementation methods: The position located at a first preset sampling distance ahead of the first vehicle's direction of travel is taken as the first sampling position. The first preset sampling distance can be 5 meters, 7 meters, etc., without specific limitation. Similarly, a second sampling position, a third sampling position, etc., can be determined. Each sampling position is located on the road ahead of the first vehicle's direction of travel and at a different distance from the first vehicle. For example, the first sampling position is 5 meters away from the first vehicle, the second sampling position is 8 meters away, the third sampling position is 10 meters away, etc.

[0150] For example, the method for determining the first sampling position in step 1602 can be the same as the method for determining the first sampling position in step 302 described above, and will not be repeated here. Similarly, the method for determining the horizontal partition corresponding to each sampling position can be the same as the method for determining the horizontal partition corresponding to each sampling position in step 1002, and will not be repeated here.

[0151] In this embodiment of the application, after determining the horizontal partition corresponding to each sampling position, the horizontal partition information of each horizontal partition in the plurality of horizontal partitions corresponding to each sampling position in at least one sampling position can be determined:

[0152] The following parameters are determined for each sampling location: width information of each lateral partition, lane type of the lane containing each lateral partition (if the lateral partition is located between two lanes, the lane type of either lane is selected as the lane type of the lateral partition, or the lane type of the lane occupying more lanes is selected as the lane type of the lateral partition), first offset distance of each lateral partition relative to the first vehicle in the lateral direction, second offset distance of each lateral partition relative to the lane centerline of the first lane, and third deflection angle of each lateral partition relative to the travel direction of the first vehicle. For example, the offset distance of each lateral partition can be calculated using the center point position of the lateral partition. The third deflection angle ranges from -180° to 180°. ○ (where -180) ○ and 180 ○ (The third deflection angle is the same). For example, the third deflection angle can be determined as follows: the angle between the lane-defined driving direction of the lane corresponding to each lateral section and the driving direction of the first vehicle is taken as the third deflection angle. If the lane-defined driving direction of the lane corresponding to each lateral section deflects to the right relative to the driving direction of the first vehicle at the current moment, then the third deflection angle is 0. ○ ~180 ○ Between; if the lane-defined travel direction of each lateral zone is deflected to the left relative to the travel direction of the first vehicle at the current moment, then the third deflection angle is -180. ○ ~0 ○ between.

[0153] Step 1603: The first vehicle plans its driving path based on the environmental information and the lateral partition information of each of the multiple lateral partitions corresponding to each sampling location.

[0154] Specifically, when planning its own travel path, the first vehicle may include, for example: Figure 17 The following steps are shown:

[0155] Step 1701: The first vehicle determines the partition prediction result of each of the multiple horizontal partitions corresponding to each sampling location based on the environmental information and the horizontal partition information of each horizontal partition in the multiple horizontal partitions corresponding to each sampling location. The partition prediction result includes a first prediction probability value, which indicates the probability that the first vehicle may pass through the horizontal partition, and a first prediction deflection angle of each horizontal partition relative to the driving direction of the first vehicle.

[0156] For example, environmental information (road topology information, traffic participant information) and the lateral partition information of each of the multiple lateral partitions corresponding to each sampling location can be input into the partition prediction model to obtain the partition prediction results of each lateral partition corresponding to each sampling location where the first vehicle will pass in the future.

[0157] In the embodiments of this application, the partition prediction model is obtained by training according to steps 201 to 203, which will not be described in detail here.

[0158] Step 1702: The first vehicle plans its driving path based on the partition prediction results of each of the multiple horizontal partitions corresponding to each sampling location.

[0159] For example, regarding the first sampling position, which can be any one of at least one sampling position, if any of the lateral partitions corresponding to the first sampling position intersects with the hard lane boundary line, then the first prediction probability value of the lateral partition intersecting with the hard lane boundary line is set to -1, meaning that lateral partitions intersecting with the hard lane boundary line are not considered subsequently. If, among the lateral partitions corresponding to the first sampling position, the lane type of the lane containing the lateral partition conflicts with the current driving direction of the first vehicle, then the first prediction probability value of the lateral partition conflicting with the current driving direction of the first vehicle is set to -1, meaning that lateral partitions conflicting with the current driving direction of the first vehicle are not considered subsequently. For example, if the lane type of the lane containing the lateral partition is a straight-ahead lane from south to north, and the current driving direction of the first vehicle is from north to south, then the lane type of the lane containing the lateral partition conflicts with the current driving direction of the first vehicle.

[0160] In step 1702 above, the first vehicle plans its driving path based on the partition prediction results of each of the multiple lateral partitions corresponding to each sampling location. This path may include, for example: Figure 18 The following steps are shown:

[0161] Step 1801: The first vehicle determines at least one target lateral partition corresponding to each sampling position based on the partition prediction results of each lateral partition among multiple lateral partitions corresponding to each sampling position.

[0162] For example, for a first sampling position, which is any one of at least one sampling position, at least one target lateral partition corresponding to the first sampling position can be determined in the following two possible ways.

[0163] The first method is to take the horizontal partition with the highest first prediction probability value among the multiple horizontal partitions corresponding to the first sampling position as the target horizontal partition corresponding to the first sampling position.

[0164] In a second possible implementation, the horizontal partition with the highest first predicted probability value among the multiple horizontal partitions corresponding to the first sampling position is taken as the first target horizontal partition corresponding to the first sampling position, and at least one horizontal partition that satisfies a preset partitioning relationship with the first target horizontal partition is taken as the second target horizontal partition. Both the first target horizontal partition and the second target horizontal partition are target horizontal partitions.

[0165] In this application embodiment, at least one horizontal partition that satisfies a preset partitioning relationship with the first target horizontal partition can be understood as at least one horizontal partition that is closest to the position of the first target horizontal partition. For example, the two horizontal partitions that are closest to the position of the first target horizontal partition are both second target horizontal partitions.

[0166] Step 1802: The first vehicle plans its driving path based on the lateral partition information of at least one target lateral partition corresponding to each sampling position.

[0167] In this embodiment of the application, based on the lateral partition information of at least one target lateral partition corresponding to each sampling position, the position information and deflection angle information of the prediction point corresponding to each sampling position are determined, and the driving path of the first vehicle is planned based on the position information and deflection angle information of the prediction point corresponding to each sampling position.

[0168] For example, regarding the first sampling position, which is any one of at least one sampling position, the location information of the predicted point corresponding to the first sampling position can be obtained through, for example... Figure 19 The following steps are shown to determine this.

[0169] Step 1901: The first vehicle determines the second position information corresponding to the intersection of the third reference line and the first reference line corresponding to the first sampling position.

[0170] Step 1902: The first vehicle determines the second lateral offset distance based on the first lateral offset distance of each of the at least one target lateral partitions relative to the third reference line. The first lateral offset distance of each of the at least one target lateral partitions relative to the third reference line can be understood as the vertical distance of the center point of each of the at least one target lateral partitions relative to the third reference line.

[0171] In this embodiment of the application, if there is a target horizontal partition, the first horizontal offset distance of the target horizontal partition relative to the third reference line is used as the second horizontal offset distance.

[0172] If there are multiple target horizontal partitions (i.e., including the first target horizontal partition and at least one second target horizontal partition), then each target horizontal partition is assigned a different weight, and the second horizontal offset distance is determined based on the first horizontal offset distance of each target horizontal partition relative to the third reference line and the weight of each target horizontal partition.

[0173] For example, multiple target horizontal partitions include a first target horizontal partition, a second target horizontal partition 1, and a second target horizontal partition 2. The first horizontal offset distance of the first target horizontal partition relative to the third reference line is e1, the first horizontal offset distance of the second target horizontal partition 1 relative to the third reference line is e2, the first horizontal offset distance of the second target horizontal partition 2 relative to the third reference line is e3, the weight corresponding to the first target horizontal partition is 0.8, the weight corresponding to the second target horizontal partition 1 is 0.1, the weight corresponding to the second target horizontal partition 2 is 0.1, and the second horizontal offset distance e4 = e1*0.8 + e2*0.1 + e3*0.1.

[0174] In this embodiment of the application, since the first lateral offset distances of multiple target lateral partitions relative to the third reference line are referenced when determining the second lateral offset distance, abrupt changes in the determined second lateral offset distance can be effectively avoided, reducing errors.

[0175] Optionally, after determining the second lateral offset distance based on the first lateral offset distance of each of the at least one target lateral partitions relative to the third reference line, at least one third lateral offset distance of the first vehicle within the second historical time period can also be determined, and the second lateral offset distance can be adjusted by using at least one third lateral offset distance. The second historical time period is a preset historical time period before the current time of the first vehicle. For example, the preset historical time period can be 5 seconds or 3 seconds, etc.

[0176] In one possible implementation, at least one second historical moment is obtained from the second historical time period, and the third lateral offset distance corresponding to the first vehicle at the first or second historical moment is determined respectively. Each third lateral offset distance and the second lateral offset distance are assigned different weights, and a new second lateral offset distance is determined based on each third lateral offset distance and the second lateral offset distance, and their respective weights.

[0177] For example, the second lateral offset distance is e4. From the second historical time period, the second historical time 1 and the second historical time 2 are obtained. The third lateral offset distance of the first vehicle at the second historical time 1 is e5, and the third lateral offset distance of the first vehicle at the second historical time 2 is e6. The weight of the third lateral offset distance e5 is 0.1, the weight of the third lateral offset distance e6 is 0.1, and the weight of the second lateral offset distance e4 is 0.8. The updated second lateral offset distance e7 = e4*0.8 + e5*0.1 + e6*0.1.

[0178] In this embodiment of the application, the second lateral offset distance is adjusted by using the third lateral offset distance corresponding to the first vehicle at at least one second historical moment, which can effectively avoid abrupt changes in the determined second lateral offset distance and reduce errors.

[0179] Step 1903: The first vehicle determines the position information of the prediction point corresponding to the first sampling position based on the second position information corresponding to the first sampling position and the second lateral offset distance.

[0180] like Figure 20 As shown, the first position of the first vehicle is point B, the second position information corresponding to the first sampling position is point O, and the second lateral offset distance is the distance between point O and point A. Therefore, the position information of the predicted point corresponding to the first sampling position can be determined, namely point A.

[0181] For the first sampling position, which is any one of at least one sampling position, the deflection angle information of the prediction point corresponding to the first sampling position may include the following steps: the first vehicle determines the deflection angle information according to the first predicted deflection angle corresponding to at least one target lateral partition corresponding to the first sampling position.

[0182] In this embodiment of the application, if there is a target horizontal partition, the first predicted deflection angle corresponding to the target horizontal partition is used as the deflection angle information.

[0183] If there are multiple target horizontal partitions (i.e., including the first target horizontal partition and at least one second target horizontal partition), the deflection angle information is determined according to the first preset deflection angle corresponding to each target horizontal partition and the weight of each target horizontal partition.

[0184] For example, multiple target horizontal partitions include a first target horizontal partition, a second target horizontal partition 1, and a second target horizontal partition 2. The first preset deflection angle corresponding to the first target horizontal partition is d1, the first preset deflection angle corresponding to the second target horizontal partition 1 is d2, and the first preset deflection angle corresponding to the second target horizontal partition 2 is d3. The weight corresponding to the first target horizontal partition is 0.8, the weight corresponding to the second target horizontal partition 1 is 0.1, and the weight corresponding to the second target horizontal partition 2 is 0.1. The deflection angle information = d1*0.8 + d2*0.1 + d3*0.1.

[0185] In this embodiment of the application, the driving path of the first vehicle is planned based on the position information of the prediction point corresponding to each sampling position and the deflection angle information. This may include the following steps: The first vehicle uses a Dubins curve to perform curve fitting on the first vehicle's current position information, the position information of the prediction point corresponding to each sampling position, and the deflection angle information to obtain the driving path of the first vehicle.

[0186] In the above method, since the width of the lateral partition is smaller than the width of the lane, the accuracy of the planned driving path of the first vehicle can be improved based on environmental information and the lateral partition information of multiple lateral partitions corresponding to each sampling position.

[0187] This application embodiment also provides a path planning device, the path planning device being suitable for... Figure 1 In the illustrated application scenario, the path planning device can optionally be a computing device with path planning functionality or a device (such as a plug-in, component, or functional part) capable of supporting the computing device in implementing the path planning method. Optionally, in one example, the path planning device is used to implement the above... Figure 2 or Figure 16 The technical solutions involved in the path planning device in the method embodiments shown, or the modules (such as chips) of the path planning device used to implement the above. Figure 2 or Figure 16 The technical solution involved in the path planning device in the method embodiment shown can therefore also achieve the above. Figure 2 or Figure 16 The beneficial effects of the path planning device in the illustrated method embodiment.

[0188] See Figure 21 The path planning device 2100 includes an acquisition module 2101 and a processing module 2102. The path planning device 2100 is used to implement the above... Figure 2 or Figure 16 The method embodiment shown illustrates the function of the path planning device.

[0189] When the path planning device 2100 is used to implement Figure 2 or Figure 16 The function of the method embodiment shown is as follows: Acquisition module 2101 is used to acquire environmental information of a first area where the first vehicle is located, wherein the first area is within the perception range of the first vehicle and includes at least one lane, and the first vehicle is located in the first lane of the at least one lane. Processing module 2102 is used to determine, based on the environmental information, the lateral partition information of each of the plurality of lateral partitions corresponding to each of the at least one sampling position, wherein each of the at least one sampling position is located ahead of the first vehicle's driving direction and at a different distance from the first vehicle, and the width of each lateral partition is less than the width of the first lane. Processing module 2102 is further used to plan the driving path of the first vehicle based on the environmental information and the lateral partition information of each of the plurality of lateral partitions corresponding to each sampling position.

[0190] Wherein, when the path planning device 2100 is used to implement Figure 2 or Figure 16 For a more detailed description of the acquisition module 2101 and the processing module 2102 regarding the function of the path planning device in the method embodiment shown, please refer to the above. Figure 2 or Figure 16 The relevant descriptions in the method embodiments shown will not be repeated here.

[0191] Optionally, the above Figure 21 Both the acquisition module and the processing module shown can be implemented in software or hardware. For example, the implementation of the processing module will be described below. Similarly, the implementation of the acquisition module can be referenced from that of the processing module, and will not be repeated here.

[0192] When implemented in software, a processing module, as an example of a software functional unit, can be an application or code block running on a computer device. The computer device can be at least one of a physical host, virtual machine, container, or other computing device. Furthermore, there can be one or more computer devices. For example, a processing module can be an application running on multiple hosts / virtual machines / containers. It should be noted that the multiple hosts / virtual machines / containers used to run the application can be distributed within the same availability zone (AZ) or in different AZs. Similarly, the multiple hosts / virtual machines / containers used to run the application can be distributed within the same region or in different regions. Typically, a region can include multiple AZs.

[0193] Similarly, multiple hosts / virtual machines / containers used to run the application can be distributed within the same Virtual Private Cloud (VPC) or across multiple VPCs. Typically, a region can include multiple VPCs, and a VPC can include multiple Availability Zones (AZs).

[0194] When implemented in hardware, the processing module, as an example of a hardware functional unit, may include at least one computing device, such as a server. Alternatively, the processing module may be implemented using an application-specific integrated circuit (ASIC) or a programmable logic device (PLD). The PLD can be a complex programmable logical device (CPLD), a field-programmable gate array (FPGA), generic array logic (GAL), or any combination thereof.

[0195] The processing module includes multiple computing devices that can be distributed within the same Availability Zone (AZ) or in different AZs. Similarly, the processing module can be distributed within the same region or in different regions. Likewise, the processing module can be distributed within the same Virtual Private Cloud (VPC) or multiple VPCs. These multiple computing devices can be any combination of computing devices such as servers, ASICs, PLDs, CPLDs, FPGAs, and GALs.

[0196] It should be noted that the steps implemented by the acquisition module and the processing module can be specified as needed. By implementing different execution steps of the path planning method through the acquisition module and the processing module, the full functionality of the path planning device can be achieved.

[0197] Furthermore, it should be noted that the module division in the embodiments of this application is illustrative and only represents one logical functional division. In actual implementation, there may be other division methods. The functional modules in the embodiments of this application can be integrated into one module, or each module can exist physically separately, or two or more modules can be integrated into one module. For example, the acquisition module and the processing module can be integrated into one module, or the acquisition module and the processing module can be the same module. The integrated unit described above can be implemented in hardware or as a software functional unit.

[0198] Based on the same concept, embodiments of this application also provide a path planning device, which is suitable for... Figure 1 In the illustrated application scenario, the path planning device is used to achieve the above. Figure 2 or Figure 16 The technical solution involved in the path planning device in the method embodiment shown can therefore also achieve the above. Figure 2 or Figure 16 The beneficial effects of the path planning device in the illustrated method embodiment are shown. See also Figure 22 The route planning device 2200 includes a processor 2202. Optionally, the route planning device further includes a memory 2203, and optionally, the route planning device 2200 also includes a transceiver 2201. The transceiver 2201, the processor 2202, and the memory 2203 are interconnected. When the route planning device 2200 is used to implement the technical solution related to the route planning device provided in the above embodiments, the transceiver 2201 can be used to implement the function of the acquisition module 2101, and the processor 2202 is used to implement the function of the processing module 2102.

[0199] Optionally, the transceiver 2201, processor 2202, and memory 2203 are interconnected via bus 2204. Bus 2204 can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 5 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.

[0200] Transceiver 2201 is used for receiving and sending data. In one example, the transceiver can be a transceiver device with integrated data transmission and reception capabilities. In another example, the transceiver can also consist of a transmitter and a receiver, wherein the transmitter is used to send data and the receiver is used to receive data.

[0201] Optionally, transceiver 2201 may include a transmitter and / or a receiver. The transmitter is used to send signals, messages, information, or data, etc. The receiver is used to receive signals, messages, information, or data, etc. For example, the transmitter sends signals, messages, information, or data under the control of processor 2202. The receiver receives signals, messages, information, or data under the control of processor 2202.

[0202] The functions of processor 2202 can be referred to the above. Figure 2 or Figure 16The description of the corresponding functions of the path planning device in the illustrated method embodiment is omitted here. The processor 2202 can be a central processing unit (CPU), a network processor (NP), or a combination of a CPU and an NP, etc. The processor 2202 may further include a hardware chip. The aforementioned hardware chip can be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The aforementioned PLD can be a complex programmable logic device (CPLD), a field-programmable gate array (FPGA), a generic array logic (GAL), or any combination thereof. When implementing the above functions, the processor 2202 can be implemented in hardware, or it can be implemented by hardware executing corresponding software.

[0203] The memory 2203 may include volatile memory, such as random access memory (RAM), and may also include non-volatile memory, such as at least one disk storage device.

[0204] The memory 2203 stores executable program instructions, and the processor 2202 executes these executable program instructions to implement the functions of the aforementioned path planning device (such as the acquisition module and the processing module), thereby realizing the path planning method provided in this application embodiment. That is, the memory 2203 stores computer program instructions for executing the path planning method.

[0205] Alternatively, the memory 2203 stores executable code, which the processor 2202 executes to implement the functions of the aforementioned path planning device (such as an acquisition module or a processing module), thereby implementing the path planning method provided in this application embodiment. That is, the memory 2203 stores computer program code for the path planning device to execute the path planning method provided in this application embodiment.

[0206] This application embodiment also provides a route planning system, which includes the aforementioned route planning device, wherein the route planning device can be used to implement the above-described... Figure 2 or Figure 16The technical solutions involved in the path planning device in the method embodiment shown are exemplary. The number of path planning devices can be one or more, and this application embodiment does not limit this.

[0207] This application also provides a vehicle that includes the path planning device or the path planning system described above.

[0208] This application also provides a computer program product, which includes computer program instructions. When the computer program or instructions are run on a computer, the computer executes the path planning method provided in the above embodiments.

[0209] This application also provides a computer-readable storage medium storing a computer program or computer program instructions, which, when executed by a computer, causes the computer to perform the path planning method provided in the above embodiments.

[0210] The storage medium can be any available medium that a computer can access. For example, but not limited to, a computer-readable medium can include RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage media or other magnetic storage devices, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer.

[0211] This application also provides a chip coupled to a memory, which is used to read computer program instructions stored in the memory to implement the path planning method provided in the above embodiments.

[0212] This application also provides a chip system including a processor for supporting a computer device in implementing the functions involved in the path planning device in the above embodiments. In one design, the chip system further includes a memory for storing necessary program instructions and / or data of the computer device. The chip system may be composed of chips or may include chips and other discrete devices.

[0213] The methods provided in this application can be implemented entirely or partially through software, hardware, firmware, or any combination thereof. When implemented in software, they can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available media may be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., high-density digital video discs (DVDs)), or semiconductor media (e.g., solid-state drives (SSDs)).

[0214] The steps of the methods described in the embodiments of this application can be directly embedded in hardware, a software unit executed by a processor, or a combination of both. The software unit can be stored in RAM, ROM, EEPROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium in the art. Exemplarily, the storage medium can be connected to the processor so that the processor can read information from the storage medium and write information to the storage medium. Optionally, the storage medium can also be integrated into the processor. The processor and the storage medium can be housed in an ASIC.

[0215] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to this application. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0216] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0217] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A path planning method, characterized in that, The method includes: Obtain environmental information of a first area where the first vehicle is located, wherein the first area is within the perception range of the first vehicle and the first area includes at least one lane, and the first vehicle is located in the first lane of the at least one lane; Based on the environmental information, the lateral partition information of each of the multiple lateral partitions corresponding to each sampling position in at least one sampling position is determined. Each sampling position in at least one sampling position is located in front of the first vehicle's driving direction and at a different distance from the first vehicle. The multiple lateral partitions corresponding to each sampling position are obtained by segmenting at least one lane at each sampling position in the lateral direction relative to the first vehicle. The width of each lateral partition is smaller than the width of the first lane. Based on the environmental information and the lateral partition information of each lateral partition corresponding to each sampling position, a partition prediction result for each lateral partition corresponding to each sampling position is determined. The partition prediction result includes a first prediction probability value and a first prediction deflection angle of each lateral partition relative to the driving direction of the first vehicle. Based on the partition prediction result of each lateral partition corresponding to each sampling position, the driving path of the first vehicle is planned.

2. The method according to claim 1, characterized in that, The lateral partition information includes a first offset distance of the lateral partition relative to the first vehicle in the lateral direction.

3. The method according to claim 1, characterized in that, The step of planning the driving path of the first vehicle based on the partition prediction results of each horizontal partition corresponding to each sampling position includes: Based on the partition prediction results of each horizontal partition corresponding to each sampling position, at least one target horizontal partition corresponding to each sampling position is determined; Based on the lateral partition information of at least one target lateral partition corresponding to each sampling position, the driving path of the first vehicle is planned.

4. The method according to claim 3, characterized in that, The step of determining at least one target horizontal partition corresponding to each sampling position based on the partition prediction result of each horizontal partition corresponding to each sampling position includes: The first horizontal partition among the multiple horizontal partitions corresponding to the first sampling position is taken as the first target horizontal partition corresponding to the first sampling position. The first sampling position is any one of the at least one sampling position. The first target horizontal partition is the horizontal partition with the highest prediction probability among the multiple horizontal partitions corresponding to the first sampling position in the partition prediction result.

5. The method according to claim 4, characterized in that, It also includes using at least one horizontal partition that satisfies a preset partitioning relationship with the first target horizontal partition as the second target horizontal partition.

6. The method according to any one of claims 3 to 5, characterized in that, The step of planning the driving path of the first vehicle based on the lateral partition information of at least one target lateral partition corresponding to each sampling position includes: Based on the lateral partitioning information of at least one target lateral partition corresponding to each sampling position, the position information and deflection angle information of the prediction point corresponding to each sampling position are determined. Based on the location information and deflection angle information of the prediction point corresponding to each sampling location, the driving path of the first vehicle is planned.

7. The method according to any one of claims 1 to 5, characterized in that, The environmental information includes road topology information and traffic participant information.

8. A path planning device, characterized in that, include: The acquisition module is used to acquire environmental information of a first area where the first vehicle is located. The first area is within the perception range of the first vehicle and includes at least one lane. The first vehicle is located in the first lane of the at least one lane. The processing module is configured to determine, based on the environmental information, the lateral partition information of each of the plurality of lateral partitions corresponding to each of the at least one sampling location, wherein each of the at least one sampling location is located in front of the first vehicle's driving direction and at a different distance from the first vehicle, and the plurality of lateral partitions corresponding to each sampling location are obtained by segmenting at least one lane at each sampling location in the lateral direction relative to the first vehicle, and the width of each lateral partition is less than the width of the first lane; The processing module is further configured to determine the partition prediction result of each horizontal partition corresponding to each sampling position based on the environmental information and the horizontal partition information of each horizontal partition among the multiple horizontal partitions corresponding to each sampling position, wherein the partition prediction result includes a first prediction probability value and a first prediction deflection angle of each horizontal partition relative to the driving direction of the first vehicle; and to plan the driving path of the first vehicle based on the partition prediction result of each horizontal partition corresponding to each sampling position.

9. A path planning device, characterized in that, The device includes a memory and a processor, the memory being used to store computer program instructions and data, and the processor being used to execute computer program instructions that call the memory to cause the path planning device to perform the method as described in any one of claims 1-7.

10. A path planning system, characterized in that, Includes the path planning device as described in claim 8 or 9.

11. A vehicle, characterized in that, The vehicle includes the path planning device as described in claim 8 or 9, or the vehicle includes the path planning system as described in claim 10.

12. A chip, characterized in that, The chip is coupled to a memory, and the chip reads computer program instructions stored in the memory to execute the method as described in any one of claims 1-7.

Images