Lane-level topology map construction method, computer device, storage medium, program product, and mobile platform

Abstract

Disclosed in the present application are a lane topology map construction method, a computer device, a storage medium, a program product, and a mobile platform. The lane topology map construction method comprises: on the basis of a traveling path and a road geometry map, determining a plurality of routing points; extracting a plurality of road sections between adjacent routing points among the plurality of routing points; and on the basis of the traveling path and lane information of the plurality of road sections, generating a lane topology map. In the present application, in view of a road geometry map, passing routing points are determined in real time by collecting a traveling path of a vehicle, and road segments between the routing points are extracted to construct a topology map. The whole process can be performed automatically without manual labeling. The cost of constructing a topology map is reduced, and the efficiency of constructing the topology map is improved.

Description

Lane-level topology map construction methods, computer equipment, storage media, software products, and mobile platforms Technical Field

[0001] This application relates to the field of autonomous driving technology, and in particular to a method for constructing lane-level topology maps, computer equipment, storage medium, program products, and mobile platform. Background Technology

[0002] Lane-level topology refers to the topological relationships between lanes, including lane connectivity: left-right neighbor relationships, preceding-following successor relationships, merging and diverging relationships, and lane connections before and after intersections. Lane-level topology is an essential element in high-precision maps, as autonomous vehicles rely on it to perform lane changes, merging and diverging, crossing intersections, and entering / exiting ramps. Current mainstream solutions involve collecting lane-level geometric information using high-precision data collection vehicles, transmitting it back to the cloud, and then manually annotating the topology. This method is time-consuming, costly, has a low update frequency, and limited practicality. Summary of the Invention

[0003] This application provides a lane-level topology map construction method, computer device, storage medium, program product, and mobile platform to at least solve one of the above-mentioned technical problems.

[0004] In a first aspect, embodiments of this application provide a method for constructing a lane topology map, including:

[0005] Multiple route points are determined based on driving trajectory and road geometry map;

[0006] Extract multiple road segments between adjacent routing points from the plurality of routing points;

[0007] A lane topology map is generated based on the driving trajectory and lane information of the multiple road segments.

[0008] In some embodiments, determining multiple routing points based on a driving trajectory and a road geometry map includes: determining a first set of intersection points on the driving trajectory based on the road geometry map;

[0009] At least the first set of intersection points is clustered to identify multiple routing points.

[0010] In some embodiments, the lane topology map construction method further includes:

[0011] Determine whether there are lane lines on both sides of the trajectory points on the driving trajectory; if not, determine that the trajectory points on the driving trajectory are intersection points, and determine the second set of intersection points;

[0012] The step of clustering at least the first set of intersection points to determine multiple routing points includes: clustering the first set of intersection points and the second set of intersection points to determine multiple routing points.

[0013] In some embodiments, generating a lane topology map based on the driving trajectory and lane information of the plurality of road segments includes:

[0014] Based on the lane information of the multiple road segments, a first topological relationship between lanes in each of the road segments is determined;

[0015] Based on the driving trajectory and the lane information of the multiple road segments, a second topological relationship between lanes of adjacent road segments is determined;

[0016] A lane-level topology map is generated based on the first and second topology relationships.

[0017] In some embodiments, the lane information of multiple road segments includes lane line information, lane start information, and lane position information in each of the road segments;

[0018] Based on the lane information of the multiple road segments, a first topological relationship between lanes in each of the road segments is determined, including:

[0019] Based on the lane information of each road segment, determine at least one of the following relationships among the lanes in each road segment: successive relationship, merging and diverging relationship, and left and right adjacent relationship.

[0020] In some embodiments, each road segment includes a first lane segment and a second lane segment; determining the successor relationship between lanes in each road segment based on lane information of each road segment includes: when the distance between the end point of the first lane segment and the start point of the second lane segment is less than a first distance threshold, and the angle between the line connecting the end point of the first lane segment and the start point of the second lane segment and the first lane segment is less than a preset angle threshold, then the second lane segment is determined to be the successor lane segment of the first lane segment.

[0021] In some embodiments, each road segment includes a first lane segment and a second lane segment; determining the merging and diverging relationship between lanes in each road segment based on lane information of each road segment includes: when there is a connecting path between the end point of the first lane segment and the start point of the second lane segment, and the connecting path is less than a second distance threshold, determining the connecting path as a diverging lane segment or a merging lane segment between the first lane segment and the second lane segment.

[0022] In some embodiments, each road segment includes a first lane segment and a second lane segment; determining the left-right adjacency relationship between lanes in each road segment based on lane information of each road segment includes:

[0023] When it is determined, based on the lane line information, that there is exactly one lane line between the first lane segment and the second lane segment, the first lane segment and the second lane segment are determined to be adjacent to each other; or

[0024] When it is determined based on the lane line information that the lane adjacent to the left of a lane line is the first lane segment and the lane adjacent to the right of a lane line is the second lane segment, the first lane segment and the second lane segment are determined to be adjacent to each other.

[0025] In some embodiments, the plurality of road segments include adjacent first road segments and second road segments; determining a second topological relationship between lanes of adjacent road segments based on the driving trajectory and lane information of the plurality of road segments includes:

[0026] Determine the road turning attributes from the first road segment to the second road segment;

[0027] Determine the lane steering attributes of the lanes in the first road segment;

[0028] Based on the road steering attribute and the lane steering attribute, connect the lanes in the first road segment with the lanes in the second road segment.

[0029] Secondly, embodiments of this application provide a computer device, including a memory, a processor, and a computer program stored in the memory, characterized in that the processor executes the computer program to implement the steps of the method described in any of the foregoing embodiments.

[0030] Thirdly, embodiments of this application provide a computer-readable storage medium storing a computer program / instructions thereon, characterized in that the computer program / instructions, when executed by a processor, implement the steps of the method described in any of the foregoing embodiments.

[0031] Fourthly, embodiments of this application provide a computer program product, including a computer program / instructions, characterized in that, when the computer program / instructions are executed by a processor, they implement the steps of the method described in any of the foregoing embodiments.

[0032] Fifthly, embodiments of this application provide a mobile platform, characterized in that it is equipped with the computer device described in any of the foregoing embodiments.

[0033] In this embodiment, road geometry maps and vehicle trajectories are combined to determine route points in real time, and road segments between these route points are extracted to construct a topology map. The entire process can be automated, requiring no manual annotation. Furthermore, the topology map construction process can be completed locally, eliminating the need to upload collected data, saving bandwidth, reducing the cost of topology map construction, and improving efficiency. Attached Figure Description

[0034] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0035] Figure 1 is a schematic diagram of a well-constructed road lane-level geometric map in related technologies;

[0036] Figure 2 is a flowchart of an embodiment of the lane topology map construction method of this application;

[0037] Figure 3 is a schematic diagram of an embodiment of the routing points and road segments in this application;

[0038] Figure 4 is a flowchart of another embodiment of the lane topology map construction method of this application;

[0039] Figure 5 is a flowchart of another embodiment of the lane topology map construction method of this application;

[0040] Figure 6 is a flowchart of another embodiment of the lane topology map construction method of this application;

[0041] Figure 7 is a schematic diagram of an embodiment of constructing the topological relationship between road segments in this application;

[0042] Figure 8 is a flowchart of another embodiment of the lane topology map construction method of this application;

[0043] Figure 9 is a schematic diagram of the structure of an embodiment of the computer device of this application. Detailed Implementation

[0044] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application. It should be noted that, unless otherwise specified, the embodiments and features in the embodiments of this application can be combined with each other.

[0045] It should also be noted that, in this document, the terms "comprising" or "including" include not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0046] This application aims to automatically and in real-time construct a road lane-level topology map based on a pre-constructed road geometry map (see Figure 1). Therefore, it is assumed that the road geometry map has already been constructed and includes lane lines, center lines, stop lines, pedestrian crossings, intersection areas, ramp entrance areas, ground arrows, and their types. Figure 1 shows a schematic diagram of a pre-constructed road lane-level geometry map in related technologies. It includes blue center lines, solid lines, dashed lines, double yellow lines, pedestrian crossings, stop lines, and median strips.

[0047] This application provides a method for constructing a lane topology map. This method can be executed by a local computer device, which can be configured on a mobile platform (e.g., a vehicle, robot, or drone). In this embodiment, the topology map construction process is completed on the local computer device, eliminating the need to upload collected data and saving bandwidth. Alternatively, the method can be executed by a cloud server, where the topology map construction process is completed on the server, reducing the performance requirements and computing resource consumption of the local computer device.

[0048] As shown in Figure 2, an embodiment of this application provides a lane topology map construction method, including:

[0049] S10. Determine multiple routing points based on the driving trajectory and road geometry map. The driving trajectory is the path taken by the vehicle; the routing points include, but are not limited to, intersections, ramp divergences, ramp mergings, and main / auxiliary road switching points.

[0050] For example, multiple intersection points are determined based on intersection elements such as stop lines, pedestrian crossings, intersection areas, and ramp entrances in the road geometry map. Then, corresponding route points Ti are determined based on these multiple intersection points. Each route point Ti stores the trajectory range {ts, te} on its driving trajectory. Each trajectory point contains a timestamp from the time of map creation, and the trajectory range {ts, te} is the driving trajectory from time ts to time te, which is the first segment of the driving trajectory.

[0051] S20. Extract multiple road segments between adjacent routing points from the multiple routing points.

[0052] For example, routing points include adjacent first and second routing points. The second routing point is a subsequent routing point of the first routing point. A first road segment precedes the first routing point, a second road segment exists between the first and second routing points, and a third road segment follows the second routing point. Figure 3 shows a schematic diagram of an embodiment of routing points and road segments in this application. It includes a blue center line, solid lines, dashed lines, double yellow lines, pedestrian crossings, stop lines, guide strips, and mapping trajectories (i.e., driving trajectories). Routing points include intersection Ti and ramp Ti+1, with ramp Ti+1 being a subsequent routing point of intersection Ti; road segments include road segment Si, road segment Si+1, and road segment Si+2.

[0053] S30. Generate a lane topology map based on the driving trajectory and the lane information of the multiple road segments.

[0054] The lane topology map includes lane segments and the connections between them. Each road segment includes two or more lane segments. These two or more lane segments are related as successor and successor or left and right neighbors. Generating the lane topology map includes generating the topological relationships between lane segments within each road segment and generating the lane segment topological relationships between adjacent road segments, which together constitute a lane-level topology map of the roads traversed by the driving trajectory. For example, the road turning attributes between adjacent first and second road segments are determined based on the driving trajectory; lane segments in the first road segment with lane turning attributes consistent with the aforementioned road turning attributes are selected, and topological connections are generated between them and lane segments in the second road segment, thereby generating lane segment topological relationships between adjacent road segments.

[0055] In this embodiment, road geometry maps and vehicle trajectories are combined to determine route points in real time, and road segments between these route points are extracted to construct a topology map. The entire process can be automated, requiring no manual annotation. This reduces the cost and improves the efficiency of topology map construction.

[0056] As shown in Figure 4, an embodiment of this application provides a lane topology map construction method. In this embodiment, multiple routing points are determined based on driving trajectories and road geometry maps, including:

[0057] S11. Determine the set of first intersection points on the driving trajectory based on the road geometry map.

[0058] For example, based on the intersection elements such as stop lines, pedestrian crossings, intersection areas, and ramp entrances contained in the road geometry map, multiple trajectory points on the driving trajectory that fall within the range of the aforementioned intersection elements are determined as the first intersection point set.

[0059] The intersection elements in the road geometry map are obtained through the perception modules of vehicles or roadside equipment. For example, during the construction of the geometry map, vehicles autonomously collect data through surround-view cameras, identify environmental information through the perception module, and use the environmental information to build the geometry map; or vehicles acquire data collected by roadside equipment or other vehicles, identify environmental information through the perception module, and use the environmental information to build the geometry map; or vehicles autonomously collect data through surround-view cameras, and simultaneously acquire data collected by roadside equipment or other vehicles, identify the acquired data through the perception module to obtain environmental information, and use the environmental information to build the geometry map. It should be noted that the above are only examples illustrating the data sources for constructing the geometry map, and this application is not limited to these.

[0060] In the above embodiments, the perception module can be implemented based on open-source models such as BEV-former, which is an algorithm model that uses pure vision for perception tasks. BEV-former extracts image features from the vehicle's surround-view cameras and transforms the extracted surround-view features into the BEV (Bird's-eye-view) space through model learning (the model learns how to transform features from the image coordinate system to the BEV coordinate system), thereby achieving 3D object detection and map segmentation tasks. This application does not limit this aspect.

[0061] S12. At least the first set of intersection points is clustered to determine multiple routing points.

[0062] For example, the first set of intersection points is clustered to obtain multiple classes of intersection points, and these multiple classes of intersection points are determined as multiple routing points. The clustering algorithm can be a general one such as DBSCAN (Density-Based Spatial Clustering of Applications with Noise), and this application does not limit its application to this method.

[0063] In this embodiment, intersection elements in the road geometry map are used to determine the trajectory points falling within the range of intersection elements on the driving trajectory as the first intersection point set. A clustering algorithm is then used to classify the first intersection point set, thereby determining multiple routing points. By using intersection elements from the pre-constructed road geometry map, the accuracy and reliability of the selected intersection elements are ensured, thus improving the accuracy and reliability of the determined routing points.

[0064] In the process of developing this application, the inventors discovered that the image recognition method using a sensing module in the aforementioned embodiments has the problem of missed detections at intersections. Therefore, the following improvement scheme is further proposed:

[0065] The embodiments of this application provide a lane topology map construction method. Based on the foregoing embodiments, the method further includes: determining whether there are lane lines on both sides of the trajectory points on the driving trajectory; if not, determining the trajectory points on the driving trajectory as intersection points to determine a second set of intersection points; the step of clustering at least the first set of intersection points to determine multiple routing points includes: clustering the first set of intersection points and the second set of intersection points to determine multiple routing points.

[0066] The determination of whether there are lane lines on both sides of the trajectory point on the driving trajectory includes: if there are no lane lines on both sides of the trajectory point, then the trajectory point is marked as an intersection point; for example, each trajectory point has x and y coordinates and a heading angle, based on which the normal direction of each trajectory point can be obtained, if the normal of the trajectory point does not intersect with the lane line, then it is determined that there are no lane lines on both sides of the trajectory point.

[0067] In this embodiment, the method for clustering the first set of intersection points and the second set of intersection points can refer to the aforementioned embodiment, and will not be repeated here.

[0068] In this embodiment, the second intersection point set is determined by judging whether there are lane lines on the left and right sides of the trajectory point. Finally, the first intersection point set and the second intersection point set are clustered to determine multiple routing points, which effectively solves the problem of missed intersection points when image recognition is performed only by the perception module.

[0069] As shown in Figure 5, an embodiment of this application provides a method for constructing a lane topology map. The method for generating a lane topology map based on the driving trajectory and lane information of the multiple road segments includes:

[0070] S31. Based on the lane information of the multiple road segments, determine the first topological relationship between lanes in each of the multiple road segments. The lane information of the multiple road segments includes lane line information, lane start-point information, lane end-point information, and lane position information for each of the road segments. Lane line information includes lane boundary line information and lane centerline information; lane start-point information includes the coordinates of the starting point of the centerline of the corresponding lane segment; lane end-point information includes the coordinates of the ending point of the centerline of the lane segment; lane position information includes the relative position information between the lane segments and lane lines within a road segment (e.g., the distribution of lane segments on the left and right sides of a certain lane line).

[0071] For example, determining a first topological relationship between lanes in each of the plurality of road segments based on lane information includes: determining at least one of the following relationships among successor-successor relationships, merging-divergence relationships, and left-right neighbor relationships among lanes in each of the road segments based on lane information of each of the road segments. Specifically, the successor-successor relationships among lanes in each of the road segments can be determined based on lane line information, lane start information, and lane end information of each of the road segments; the merging-divergence relationships among lanes in each of the road segments can be determined based on lane line information, lane start information, and lane end information of each of the road segments; and the left-right neighbor relationships among lanes in each of the road segments can be determined based on lane position information of each of the road segments.

[0072] S32. Based on the driving trajectory and the lane information of the multiple road segments, determine the second topological relationship between lanes of adjacent road segments. The relationship between adjacent road segments includes routing points.

[0073] Each road segment may include multiple lanes, and each lane may form at least one lane segment. For example, the multiple road segments include a first road segment and a second road segment located on both sides of an intersection area. The second road segment is a subsequent road segment of the first road segment. Both the first and second road segments include multiple lane segments, where the multiple lane segments in the first road segment include entrance lane segments, and the multiple lane segments in the second road segment are exit lane segments. The entrance lane segments in the first road segment can be determined based on the driving trajectory, thereby establishing a topological relationship between the entrance lane segments of the first road segment and the exit lane segments of the second road segment, which is used for a second topological relationship.

[0074] S33. Generate a lane-level topology map based on the first and second topology relationships. In this embodiment, the first topology relationship between lanes within each road segment is first determined step by step, and the second topology relationship between lanes between each road segment is further determined. The first and second topology relationships are then used as the lane-level topology map. This avoids the formation of intersecting and confusing topology relationships between lane segments within a road segment and between lane segments crossing road segments.

[0075] In some embodiments, the following are examples of specific implementation methods for determining at least one of the following relationships among lanes in each road segment: successive relationship, merging / diverging relationship, and left / right neighbor relationship, based on lane information of each road segment:

[0076] In some embodiments, each road segment includes a first lane segment and a second lane segment; determining the successor relationship between lanes in each road segment based on lane information of each road segment includes: when the distance between the end point of the first lane segment and the start point of the second lane segment is less than a first distance threshold, and the angle between the line connecting the end point of the first lane segment and the start point of the second lane segment and the first lane segment is less than a preset angle threshold, then the second lane segment is determined to be the successor lane segment of the first lane segment.

[0077] For example, for any lanes Li and Lj, let the endpoint of lane Li be pi and the starting point of lane Lj be pj. If the distance between the endpoint pi and the starting point pj is distance(pi,pj) < 10m and the angle between the line connecting the endpoint pi and the starting point pj and the centerline of lane Li is angle_distance(pi,pj) < 30°, then the successor of lane Li is lane Lj, ​​and the predecessor of lane Lj is lane Li. Here, the endpoint pi and the starting point pj are points on the centerlines of the corresponding lanes Li and Lj, respectively.

[0078] In some embodiments, each road segment includes a first lane segment and a second lane segment; determining the merging and diverging relationship between lanes in each road segment based on lane information of each road segment includes: when there is a connecting path between the end point of the first lane segment and the start point of the second lane segment, and the connecting path is less than a second distance threshold, determining the connecting path as a diverging lane segment or a merging lane segment between the first lane segment and the second lane segment.

[0079] For example, for any lane Li and lane Lj, ​​if there exists a collision-free (i.e., non-collision with road edges, safety islands, guide strips, etc., i.e., non-intersection) path between the end point pi of lane Li and the starting point pj of lane Lj, ​​which does not pass through lane lines and has a shortest connecting path length of <30m (based on experience, the length of merging and diverging is generally less than 30m), then a virtual lane is generated from pi to pj as the merging and diverging lane Lk, and the successor of Li is Lk, the predecessor of Lk is Li, the successor of Lk is Lj, and the predecessor of Lj is Lk.

[0080] The method for finding connecting paths is not limited; classic algorithms such as Dijstra's algorithm or A* can be used. For example, a two-dimensional grid map can be generated using lane lines and median strips, with a value of 1 for occupied grid cells and 0 for unoccupied cells. Then, the A* algorithm can be used to search for the optimal path (shortest connecting path) in this two-dimensional grid map.

[0081] The method for generating merging and diverging lanes is not limited. Collision-free smooth lanes can be obtained by using the aforementioned searched connecting paths and optimizing them. The optimization method can be, for example, one of TEB (Time-Elastic Band, a local path optimization algorithm), MPC (Model Predictive Control), or iLQR (Iterative Linear Quadratic Regulator). This application does not limit the specific method used.

[0082] In some embodiments, each road segment includes a first lane segment and a second lane segment; determining the left-right adjacency relationship between lanes in each road segment based on lane information of each road segment includes: when it is determined based on the lane position information that there is one and only one lane line between the first lane segment and the second lane segment, determining that the first lane segment and the second lane segment are left-right adjacency; or, when it is determined based on the lane line information that the lane adjacent to the left of a lane line is the first lane segment and the lane adjacent to the right of a lane line is the second lane segment, determining that the first lane segment and the second lane segment are left-right adjacency.

[0083] For example, for any lane line, if its left nearest neighbor lane is Li and its right nearest neighbor lane is Lj, then Li and Lj can be determined to be left and right adjacent lanes to each other.

[0084] As shown in Figure 6, an embodiment of this application provides a lane topology map construction method. In this embodiment, multiple road segments include adjacent first road segment Si and second road segment Si+1. Determining the second topological relationship between lanes of adjacent road segments based on the driving trajectory and lane information of the multiple road segments includes:

[0085] S321. Determine the road turning attributes from the first road segment Si to the second road segment Si+1. The road turning attributes refer to whether the vehicle will turn left, right, make a U-turn, or go straight at the intersection. The road turning attributes can be determined based on the vehicle's trajectory.

[0086] For example, to determine the road turning attribute from Si to Si+1: Let the angle difference between the trajectory at the end of Si and the trajectory at the beginning of Si+1 be n°. If |n°| <= 30°, the road turning attribute is straight; if 145° > |n°| > 30°, the road turning attribute is left turn; if -145° < |n°| < -30°, the road turning attribute is right turn; if |n| >= 145°, the road turning attribute is U-turn.

[0087] S322. Determine the lane turning attributes of the lanes in the first road segment. The lane turning attribute refers to the presence of a ground arrow at the entrance indicating whether the lane is for straight, left, or right turns. Lane turning attributes can be determined based on image recognition.

[0088] For example, to determine the turning attribute of the entrance lane in the lane segment of the first road segment Si: find the ground arrow Mj that is closest to each entrance lane Li (obtained through image recognition) based on the distance, and the turning type of the entrance lane Li is the type of Mj.

[0089] S323. Based on the road turning attribute and the lane turning attribute, connect the lanes in the first road segment with the lanes in the second road segment.

[0090] For example, to connect the set of entrance lanes Li{Li1,Li2…Lin} of the first road segment Si with the set of exit lanes Lj{Lj1,Lj2…Ljm} of the second road segment Si+1: select lanes in the entrance lanes of Si whose lane turning attributes are the same as the road turning attributes, and set them as set Li'{Li1',Li2'…Lik'}. Then, generate a connecting lane between each pair of Li' and Lj. The predecessor of this connecting lane is Lii', and the successor is Ljj (wherein, one entrance lane in Li' is connected to at least one exit lane in Lj). As shown in Figure 7, the set Li' is {Li1,Li2,Li3}, and the set Lj is {Lj1,Lj2,Lj3}. In this process, a connecting lane Lij1 is generated between lanes Li1 and Lj1. The predecessor of connecting lane Lij1 is Li1, and its successor is Lj1. Similarly, the predecessor of connecting lane Lij2 is Li2, and its successor is Lj2; the predecessor of connecting lane Lij3 is Li3, and its successor is Lj3. Furthermore, if the lane turning attribute of lane Li1 is left turn, which is inconsistent with the road turning attribute (straight), the connecting lane from Li1 to Lj1 will not be generated. Instead, a connecting lane from Li2 to Lj1, and also from Li2 to Lj2, can be generated. The method for generating connecting lanes is not limited; they can be generated using spline curves such as Bézier curves, or collision-free curves can be generated through optimization methods, such as the iLQR method.

[0091] In this embodiment, lane selection is performed based on the similarity between lane turning attributes and road turning attributes, and topological connections are established to ensure the accuracy and effectiveness of topological connections. For example, in the entrance lane set Li{Li1, Li2, Li3} of the first road segment Si, where Li1 has a left turn lane turning attribute, Li2 has a straight lane turning attribute, and Li3 has a straight / right turn lane turning attribute, and the road turning attribute is determined to be straight based on the driving trajectory, then the set of lanes in the entrance lanes of road segment Si with the same lane turning attribute as the road turning attribute is selected as Li'{Li2, Li3}, thereby avoiding the topological connection (erroneous topological connection) between the left turn lane Li1 and the exit lane of road segment Si+1.

[0092] As shown in Figure 8, an embodiment of this application provides a method for constructing a lane topology map.

[0093] 1. Obtain a geometric map, collect data using cameras on the data acquisition vehicle, and generate a driving trajectory.

[0094] II. Route Point Identification: Identifying road route points (including but not limited to intersections, ramp divergences, ramp mergings, and main / auxiliary road switching points). Each route point Ti stores its trajectory range {ts, te}. Each trajectory point contains a timestamp from the time of mapping. This trajectory range {ts, te} represents the first segment of the driving trajectory from time ts to time te. Specific method:

[0095] 1. If there are no lane lines on either side of a trajectory point, then mark the trajectory point as an intersection point; for example, each trajectory point has x and y coordinates and a heading angle. Based on this, the normal direction of each trajectory can be obtained. If the normal does not intersect with the lane lines, then it is determined that there are no lane lines on either side of the trajectory point.

[0096] 2. Mark trajectory points within the intersection element range as intersection points. Intersection elements include, but are not limited to: stop lines, pedestrian crossings, intersection areas, and ramp entrances. Intersection elements can be identified through a vehicle perception module or obtained through a geometric map.

[0097] 3. Cluster all intersection points based on distance, with each cluster ultimately representing a routing point, denoted as T1, T2…Tn. For example, the general DBSCAN clustering algorithm can be used. Since the distance between two intersections is typically large, setting a reasonable threshold is sufficient to ensure accurate clustering. The specific threshold can be determined based on the actual distance between the two intersections; this application does not impose any limitations on this.

[0098] 3. Road Segment Extraction: Extract the lanes between two consecutive routing points Ti and Ti+1, which constitute a road segment without traffic diversion or branching, and denote it as a basic unit Si. See Figure 3 for reference.

[0099] IV. Perform road topology inference for each road segment Si, using the following method:

[0100] 1. Road segment successor inference: For any lanes Li and Lj, let the endpoint of Li be pi and the starting point of Lj be pj. If distance(pi,pj) < 10m && angle_distance(pi,pj) < 30°, then the successor of Li is Lj, and the predecessor of Lj is Li. Here, endpoint pi and starting point pj are points on the centerlines of the corresponding lanes Li and Lj.

[0101] 2. Road segment merging and diverging inference: For any lanes Li and Lj, if there exists a collision-free (i.e., non-collision with curbs, safety islands, or guide strips, i.e., non-intersection) path between the endpoint pi of Li and the starting point pj of Lj, which does not pass through lane lines, and the shortest connecting path length is <30m (based on experience, the merging and diverging length is generally less than 30m), then a virtual lane is generated from pi to pj as the merging and diverging lane Lk, and the successor of Li is Lk, the predecessor of Lk is Li, the successor of Lk is Lj, and the predecessor of Lj is Lk.

[0102] 3. Road segment left and right neighbor inference: For any lane line, if its left nearest neighbor is Li and its right nearest neighbor is Lj, then Li and Lj are left and right neighbors.

[0103] V. Inter-road segment connection: Connect two consecutive road segments at the lane level according to the route point type.

[0104] 1. Based on the aforementioned steps, the lane-level topology of each road segment can be obtained.

[0105] 2. Determine the road turning attribute from the first road segment Si to the second road segment Si+1: Let the angle difference between the trajectory at the end of Si and the trajectory at the beginning of Si+1 be n°. If |n°| <= 30°, then it is going straight; if 145° > |n°| > 30°, then it is turning left; if -145° < |n°| < -30°, then it is turning right; if |n°| >= 145°, then it is making a U-turn.

[0106] 3. Determine the turning type attribute of the entrance lane: Find the ground arrow Mj that is closest to each entrance lane Li (obtained through image recognition) based on the distance. Then the turning type of the entrance lane Li is the type of Mj.

[0107] 4. Connect the entrance lane set Li{Li1,Li2…Lin} of road segment Si with the exit lane set Lj{Lj1,Lj2…Ljm} of road segment Si+1: Select lanes in the entrance lanes of Si whose lane turning attributes are the same as the road turning attributes, and set them as Li'{Li1',Li2'…Lik'}. Then, generate a connecting lane between each pair of Li' and Lj. The predecessor of this connecting lane is Lii', and the successor is Ljj. As shown in Figure 7, a connecting lane Lij1 is generated between lanes Li1 and Lj1. The predecessor of the connecting lane Lij1 is Li1, and the successor is Lj1. The method for generating the connecting lane is not limited. It can be generated using spline curves such as Bézier curves, or it can be generated by optimization to generate collision-free curves, such as the iLQR method.

[0108] Traditional methods require manual topology annotation, which is time-consuming and labor-intensive. The topology mapping method disclosed in this application can automatically generate a topology map while the data collection vehicle is collecting data, requiring no manual intervention and significantly reducing annotation costs. Traditional methods require uploading large amounts of sensor data to the cloud for manual topology annotation, resulting in high data usage and storage costs. The method disclosed in this application can automatically build the topology map on the vehicle itself, so only the topology map needs to be uploaded to the cloud. This requires a much smaller data volume, significantly reducing data usage costs and saving overall expenses.

[0109] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of combined actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Secondly, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application. In the above embodiments, the descriptions of each embodiment have their own emphasis; for parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0110] In some embodiments, this application also provides a lane topology map building apparatus, which can be configured on a mobile platform (e.g., a vehicle, robot, and drone). In this embodiment, the topology map building process is completed on a local computer device, eliminating the need to upload collected data and saving bandwidth. The method can also be executed by a cloud server, with the topology map building process completed on the server, reducing the performance requirements and computing resource consumption of the local computer device. The lane topology map building apparatus includes:

[0111] The route point determination module is used to determine multiple route points based on the vehicle trajectory and road geometry map;

[0112] The road segment extraction module is used to extract multiple road segments between adjacent routing points from the multiple routing points.

[0113] The map generation module is used to generate a lane topology map based on the driving trajectory and lane information of the multiple road segments.

[0114] In this embodiment, road geometry maps and vehicle trajectories are combined to determine route points in real time, and road segments between these route points are extracted to construct a topology map. The entire process can be automated, requiring no manual annotation. This reduces the cost and improves the efficiency of topology map construction.

[0115] In some embodiments, determining multiple routing points based on a driving trajectory and a road geometry map includes: determining a first set of intersection points on the driving trajectory based on the road geometry map;

[0116] At least the first set of intersection points is clustered to identify multiple routing points.

[0117] In some embodiments, the lane topology map construction device further includes: an intersection point determination module, used to determine whether there are lane lines on both sides of the trajectory point on the driving trajectory; if not, the trajectory point on the driving trajectory is determined as an intersection point, so as to determine a second intersection point set; the step of clustering at least the first intersection point set to determine multiple routing points includes: clustering the first intersection point set and the second intersection point set to determine multiple routing points.

[0118] In some embodiments, generating a lane topology map based on the driving trajectory and lane information of the plurality of road segments includes:

[0119] Based on the lane information of the multiple road segments, a first topological relationship between lanes in each of the road segments is determined;

[0120] Based on the driving trajectory and the lane information of the multiple road segments, a second topological relationship between lanes of adjacent road segments is determined;

[0121] A lane-level topology map is generated based on the first and second topology relationships.

[0122] In some embodiments, the lane information of multiple road segments includes lane line information, lane start information, lane end information and lane position information in each of the road segments;

[0123] Based on the lane information of the multiple road segments, a first topological relationship between lanes in each of the road segments is determined, including:

[0124] Based on the lane information of each road segment, at least one of the following relationships among the lanes in each road segment is determined: successive relationship, merging / diverging relationship, and left / right adjacency relationship. Specifically, the successive relationship among lanes in each road segment can be determined based on the lane line information, lane start information, and lane end information of each road segment; the merging / diverging relationship among lanes in each road segment can be determined based on the lane line information, lane start information, and lane end information of each road segment; and the left / right adjacency relationship among lanes in each road segment can be determined based on the lane position information of each road segment.

[0125] In some embodiments, each road segment includes a first lane segment and a second lane segment; determining the successor relationship between lanes in each road segment based on lane information of each road segment includes: when the distance between the end point of the first lane segment and the start point of the second lane segment is less than a first distance threshold, and the angle between the line connecting the end point of the first lane segment and the start point of the second lane segment and the first lane segment is less than a preset angle threshold, then the second lane segment is determined to be the successor lane segment of the first lane segment.

[0126] In some embodiments, each road segment includes a first lane segment and a second lane segment; determining the merging and diverging relationship between lanes in each road segment based on lane information of each road segment includes: when there is a connecting path between the end point of the first lane segment and the start point of the second lane segment, and the connecting path is less than a second distance threshold, determining the connecting path as a diverging lane segment or a merging lane segment between the first lane segment and the second lane segment.

[0127] In some embodiments, each road segment includes a first lane segment and a second lane segment; determining the left-right adjacency relationship between lanes in each road segment based on lane information of each road segment includes:

[0128] When it is determined, based on the lane line information, that there is exactly one lane line between the first lane segment and the second lane segment, the first lane segment and the second lane segment are determined to be adjacent to each other; or

[0129] When it is determined based on the lane line information that the lane adjacent to the left of a lane line is the first lane segment and the lane adjacent to the right of a lane line is the second lane segment, the first lane segment and the second lane segment are determined to be adjacent to each other.

[0130] In some embodiments, the plurality of road segments include adjacent first road segments and second road segments; determining a second topological relationship between lanes of adjacent road segments based on the driving trajectory and lane information of the plurality of road segments includes:

[0131] Determine the road turning attributes from the first road segment to the second road segment;

[0132] Determine the lane steering attributes of the lanes in the first road segment;

[0133] Based on the road steering attribute and the lane steering attribute, connect the lanes in the first road segment with the lanes in the second road segment.

[0134] In some embodiments, this application provides a computer device including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the method described in any of the foregoing embodiments.

[0135] In some embodiments, this application provides a computer-readable storage medium storing a computer program / instructions thereon, characterized in that the computer program / instructions, when executed by a processor, implement the steps of the method described in any of the foregoing embodiments.

[0136] In some embodiments, this application provides a computer program product including a computer program / instructions, characterized in that the computer program / instructions, when executed by a processor, implement the steps of the method described in any of the foregoing embodiments.

[0137] In some embodiments, this application provides a mobile platform characterized by having a computer device as described in any of the foregoing embodiments installed. The mobile platform may be a vehicle, robot, drone, etc.

[0138] Figure 9 is a schematic diagram of the hardware structure of a computer device for performing a lane topology map construction method according to another embodiment of this application. As shown in Figure 9, the device includes:

[0139] One or more processors 910 and memory 920, with one processor 910 as an example in Figure 9.

[0140] The device for performing the lane topology map construction method may also include an input device 930 and an output device 940.

[0141] The processor 910, memory 920, input device 930 and output device 940 can be connected by a bus or other means. Figure 9 shows an example of connection by bus.

[0142] The memory 920, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer-executable programs, and modules, such as the program instructions / modules corresponding to the lane topology map construction method in the embodiments of this application. The processor 910 executes various functional applications and data processing of the server by running the non-volatile software programs, instructions, and modules stored in the memory 920, thereby implementing the lane topology map construction method of the above-described method embodiments.

[0143] The memory 920 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the lane topology map building device. Furthermore, the memory 920 may include high-speed random access memory and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory 920 may optionally include memory remotely located relative to the processor 910, and this remote memory may be connected to the lane topology map building device via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0144] The input device 930 can receive input digital or character information and generate signals related to user settings and function control of the lane topology map building device. The output device 940 may include a display device such as a display screen.

[0145] The one or more modules are stored in the memory 920, and when executed by the one or more processors 910, they execute the lane topology map construction method in any of the above method embodiments.

[0146] The above-described product can perform the methods provided in the embodiments of this application, and has the corresponding functional modules and beneficial effects for performing the methods. Technical details not described in detail in this embodiment can be found in the methods provided in the embodiments of this application.

[0147] The computer device in this application embodiment exists in various forms, including but not limited to:

[0148] (1) Mobile communication devices: These devices are characterized by their mobile communication capabilities and primarily aim to provide voice and data communication. These terminals include: smartphones (e.g., iPhones), multimedia phones, feature phones, and low-end phones, etc.

[0149] (2) Ultra-mobile personal computer devices: These devices fall under the category of personal computers, possessing computing and processing capabilities, and generally also have mobile internet access features. These terminals include PDAs, MIDs, and UMPCs, such as the iPad.

[0150] (3) Portable entertainment devices: These devices can display and play multimedia content. This category includes: audio and video players (such as iPods), handheld game consoles, e-books, as well as smart toys and portable car navigation devices.

[0151] (4) Server: A device that provides computing services. The components of a server include a processor, hard disk, memory, system bus, etc. Servers are similar to general computer architectures, but because they need to provide highly reliable services, they have higher requirements in terms of processing power, stability, reliability, security, scalability, and manageability.

[0152] (5) Other electronic devices with data interaction functions.

[0153] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0154] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented using software plus a general-purpose hardware platform, or of course, using hardware. Based on this understanding, the above technical solutions, in essence or the parts that contribute to the related technology, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0155] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for constructing a lane topology map, comprising: Multiple route points are determined based on driving trajectory and road geometry map; Extract multiple road segments between adjacent routing points from the plurality of routing points; A lane topology map is generated based on the driving trajectory and lane information of the multiple road segments.

2. The method according to claim 1, characterized in that, Multiple route points are determined based on the driving trajectory and road geometry map, including: The set of first intersection points on the driving trajectory is determined based on the road geometry map; At least the first set of intersection points is clustered to identify multiple routing points.

3. The method according to claim 2, characterized in that, Also includes: Determine whether there are lane lines on both sides of the trajectory point on the driving trajectory; If it does not exist, then the trajectory points on the driving trajectory are determined as intersection points, so as to determine the second set of intersection points; The step of clustering at least the first set of intersection points to determine multiple routing points includes: clustering the first set of intersection points and the second set of intersection points to determine multiple routing points.

4. The method according to claim 1, characterized in that, A lane topology map is generated based on the driving trajectory and lane information of the multiple road segments, including: Based on the lane information of the multiple road segments, a first topological relationship between lanes in each of the road segments is determined; Based on the driving trajectory and the lane information of the multiple road segments, a second topological relationship between lanes of adjacent road segments is determined; A lane-level topology map is generated based on the first and second topology relationships.

5. The method according to claim 4, characterized in that, The lane information of the multiple road segments includes lane line information, lane start information, and lane position information in each of the road segments; Based on the lane information of the multiple road segments, a first topological relationship between lanes in each of the road segments is determined, including: Based on the lane information of each road segment, determine at least one of the following relationships among the lanes in each road segment: successive relationship, merging and diverging relationship, and left and right adjacent relationship.

6. The method according to claim 4, characterized in that, The plurality of road segments includes adjacent first road segments and second road segments; Based on the driving trajectory and lane information of the multiple road segments, a second topological relationship between lanes of adjacent road segments is determined, including: Determine the road turning attributes from the first road segment to the second road segment; Determine the lane steering attributes of the lanes in the first road segment; Based on the road steering attribute and the lane steering attribute, connect the lanes in the first road segment with the lanes in the second road segment.

7. A lane topology map construction device includes: The route point determination module is used to determine multiple route points based on the vehicle trajectory and road geometry map; The road segment extraction module is used to extract multiple road segments between adjacent routing points among the multiple routing points; The map generation module is used to generate a lane topology map based on the driving trajectory and lane information of the multiple road segments.

8. The apparatus according to claim 7, characterized in that, Multiple route points are determined based on the driving trajectory and road geometry map, including: The set of first intersection points on the driving trajectory is determined based on the road geometry map; At least the first set of intersection points is clustered to identify multiple routing points.

9. The apparatus according to claim 8, characterized in that, Also includes: The intersection point determination module is used to determine whether there are lane lines on both sides of the trajectory point on the driving trajectory; If it does not exist, then the trajectory points on the driving trajectory are determined as intersection points, so as to determine the second set of intersection points; The step of clustering at least the first set of intersection points to determine multiple routing points includes: clustering the first set of intersection points and the second set of intersection points to determine multiple routing points.

10. The apparatus according to claim 7, characterized in that, A lane topology map is generated based on the driving trajectory and lane information of the multiple road segments, including: Based on the lane information of the multiple road segments, a first topological relationship between lanes in each of the road segments is determined; Based on the driving trajectory and the lane information of the multiple road segments, a second topological relationship between lanes of adjacent road segments is determined; A lane-level topology map is generated based on the first and second topology relationships.

11. The apparatus according to claim 10, characterized in that, The lane information of the multiple road segments includes lane line information, lane start information, and lane position information in each of the road segments; Based on the lane information of the multiple road segments, a first topological relationship between lanes in each of the road segments is determined, including: Based on the lane information of each road segment, determine at least one of the following relationships among the lanes in each road segment: successive relationship, merging and diverging relationship, and left and right adjacent relationship.

12. The apparatus according to claim 10, characterized in that, The plurality of road segments includes adjacent first road segments and second road segments; Based on the driving trajectory and lane information of the multiple road segments, a second topological relationship between lanes of adjacent road segments is determined, including: Determine the road turning attributes from the first road segment to the second road segment; Determine the lane steering attributes of the lanes in the first road segment; Based on the road steering attribute and the lane steering attribute, connect the lanes in the first road segment with the lanes in the second road segment.

13. A computer device comprising a memory, a processor, and a computer program stored in the memory, characterized in that, The processor executes the computer program to implement the steps of the method according to any one of claims 1-6.

14. A computer-readable storage medium having a computer program / instructions stored thereon, characterized in that, When the computer program / instructions are executed by the processor, they implement the steps of the method described in any one of claims 1-6.

15. A computer program product comprising a computer program / instructions, characterized in that, When the computer program / instructions are executed by the processor, they implement the steps of the method described in any one of claims 1-6.

16. A mobile platform, characterized in that, The computer device of claim 13 is installed.

Images