Method and apparatus for constructing road network topology, vehicle, and storage medium

Abstract

The present application relates to a method and apparatus for constructing a road network topology, a vehicle, and a storage medium. The method comprises: acquiring a constructed road and acquiring a lane center line in the road, wherein the road comprises a predecessor road and a successor road; acquiring a learning trajectory; determining an exit lane center line of the learning trajectory on the predecessor road, and determining an entrance lane center line of the learning trajectory on the successor road; on the basis of the exit lane center line and the entrance lane center line and with reference to the learning trajectory, generating a topology connection line between the preceding road and the succeeding road; and on the basis of the topology connection line, obtaining a road network topology between the predecessor road and the successor road. The solution provided in the present application can improve the construction quality of road network topologies, thereby improving the safety of autonomous driving and the experience of autonomous driving.

Description

Road network topology construction method and device, vehicle and storage medium

[0001] The present application claims priority to the Chinese patent application No. 2024107966755, filed on June 19, 2024, and entitled "Road network topology construction method and device, vehicle and storage medium", the whole content of which is incorporated herein by reference. TECHNICAL FIELD

[0002] The present application relates to the field of autonomous driving, in particular to a road network topology construction method and device, vehicle and storage medium. BACKGROUND

[0003] Autonomous driving is a technology that uses various sensors, computer vision, artificial intelligence and machine learning to perceive, analyze and make decisions about the road environment, so as to realize the autonomous navigation and control of vehicles.

[0004] With the continuous development of autonomous driving technology, AI (Artificial Intelligence) driving has emerged. AI driving generally refers to the automatic driving of vehicles in commuting mode. AI driving is based on strong autonomous learning ability and precise route memory when the user manually drives, and can record and generate learning trajectories through one-time learning, and then execute autonomous driving according to the learning trajectories. AI driving can not depend on high-precision maps and is not limited by ODD (Operational Design Domain).

[0005] Road network topology, i.e., the connectivity of roads in space, i.e., the topological relationship, is crucial for route planning and navigation in autonomous driving. The construction quality of road network topology also affects the safety and experience of autonomous driving. Therefore, how to improve the construction quality of road network topology is a problem to be solved in autonomous driving technology. SUMMARY

[0006] To solve or partially solve the problems in the related art, the present application provides a road network topology construction method and device, vehicle and storage medium, which can improve the construction quality of road network topology and improve the safety and experience of autonomous driving.

[0007] The first aspect of the present application provides a road network topology construction method, comprising:

[0008] obtaining a constructed road and obtaining a lane center line in the road, wherein the road comprises a predecessor road and a successor road;

[0009] obtaining a learning trajectory;

[0010] determining an exit lane centerline of the predecessor road on which the learning trajectory is located, and determining an entrance lane centerline of the successor road on which the learning trajectory is located;

[0011] generating a topological connection line of the predecessor road and the successor road according to the exit lane centerline and the entrance lane centerline, and referring to the learning trajectory;

[0012] obtaining a road network topology between the predecessor road and the successor road according to the topological connection line.

[0013] In an embodiment, the generating the topological connection line of the predecessor road and the successor road according to the exit lane centerline and the entrance lane centerline, and referring to the learning trajectory, comprises:

[0014] constructing a first topological relationship between the exit lane centerline and the entrance lane centerline, and generating a first topological connection line of the predecessor road and the successor road referring to the learning trajectory;

[0015] constructing a second topological relationship between lane centerlines of the predecessor road except the exit lane centerline and lane centerlines of the successor road except the entrance lane centerline, and generating a second topological connection line of the predecessor road and the successor road referring to the learning trajectory;

[0016] The obtaining the road network topology between the predecessor road and the successor road according to the topological connection line, comprises:

[0017] obtaining the road network topology between the predecessor road and the successor road according to the first topological connection line and the second topological connection line.

[0018] In an embodiment, the predecessor road comprises a predecessor lane centerline, and the determining the exit lane centerline of the predecessor road on which the learning trajectory is located, comprises:

[0019] screening, in the predecessor lane centerline, a predecessor lane centerline with a distance to an exit end face of the predecessor road less than a first threshold value;

[0020] projecting, with the learning trajectory as a reference line, an end point of the screened predecessor lane centerline to the learning trajectory to obtain a first lateral distance value to the learning trajectory after projection;

[0021] taking the predecessor lane centerline with the smallest first lateral distance value as the exit lane centerline of the predecessor road on which the learning trajectory is located.

[0022] In an embodiment, the subsequent road comprises a subsequent lane centerline, and the determining the learning trajectory at the entry lane centerline of the subsequent road comprises:

[0023] In the subsequent lane centerline, a subsequent lane centerline with a distance to an entry end surface of the subsequent road less than a second threshold value is screened out;

[0024] A starting point of the screened-out subsequent lane centerline is projected to the learning trajectory as a reference line to obtain a second lateral distance value to the learning trajectory after the projection;

[0025] The subsequent lane centerline with the minimum second lateral distance value is taken as the learning trajectory at the entry lane centerline of the subsequent road.

[0026] In an embodiment, the constructing a second topological relationship between the lane centerline of the predecessor road except the exit lane centerline and the lane centerline of the subsequent road except the entry lane centerline comprises:

[0027] The lane centerline of the predecessor road except the exit lane centerline is projected to the learning trajectory to obtain a third lateral distance value to the learning trajectory after the projection;

[0028] The lane centerline of the subsequent road except the entry lane centerline is projected to the learning trajectory to obtain a fourth lateral distance value to the learning trajectory after the projection;

[0029] The two lane centerlines corresponding to the minimum deviation between the third lateral distance value and the fourth lateral distance value are taken to construct the second topological relationship.

[0030] In an embodiment, the generating a topological connection line between the predecessor road and the subsequent road according to the exit lane centerline and the entry lane centerline and referring to the learning trajectory comprises:

[0031] According to a preset smoothing algorithm, a smoothing trajectory is generated as the topological connection line between the predecessor road and the subsequent road from an ending point of the exit lane centerline to a starting point of the entry lane centerline with the learning trajectory as a reference.

[0032] In an embodiment, the predecessor road is the first road appearing in front of the ego vehicle, and the subsequent road is the road behind the predecessor road.

[0033] In an embodiment, the preset smoothing algorithm comprises a smoothing algorithm based on curve interpolation or a sampling-based A-star algorithm.

[0034] In an embodiment, the constructed road is constructed by the cloud or the vehicle, or constructed by the cloud and the vehicle.

[0035] In an embodiment, after the topology connection line is obtained, the road network topology between the preceding road and the succeeding road is obtained according to the topology connection line.

[0036] The topology connection line is selected as a driving path for reference in autonomous driving; or,

[0037] The driving path is regenerated according to the topology connection line for reference in autonomous driving.

[0038] The second aspect of the present application provides a road network topology construction device, comprising:

[0039] A first acquisition module is configured to acquire a constructed road and acquire a lane center line in the road, wherein the road comprises a preceding road and a succeeding road.

[0040] A second acquisition module is configured to acquire a learning trajectory.

[0041] A first processing module is configured to determine an exit lane center line of the learning trajectory on the preceding road and determine an entry lane center line of the learning trajectory on the succeeding road.

[0042] A second processing module is configured to generate a topology connection line of the preceding road and the succeeding road according to the exit lane center line and the entry lane center line and refer to the learning trajectory.

[0043] A result generation module is configured to obtain a road network topology between the preceding road and the succeeding road according to the topology connection line.

[0044] In an embodiment, the second processing module comprises:

[0045] A first topology processing submodule is configured to construct a first topology relationship between the exit lane center line and the entry lane center line and generate a first topology connection line of the preceding road and the succeeding road by referring to the learning trajectory.

[0046] A second topology processing submodule is configured to construct a second topology relationship between the lane center line of the preceding road except the exit lane center line and the lane center line of the succeeding road except the entry lane center line and generate a second topology connection line of the preceding road and the succeeding road by referring to the learning trajectory.

[0047] The result generation module obtains the road network topology between the preceding road and the succeeding road according to the first topology connection line and the second topology connection line.

[0048] In an embodiment, the first processing module comprises:

[0049] An exit lane centerline determination sub-module is configured to: in the predecessor lane centerlines, filter a predecessor lane centerline with a distance to an exit end face of the predecessor road less than a first threshold value; project an end point of the filtered predecessor lane centerline to the learned trajectory as a reference line to obtain a first lateral distance value to the learned trajectory after projection; and determine the predecessor lane centerline with the smallest first lateral distance value as the learned trajectory on the exit lane centerline of the predecessor road.

[0050] In an embodiment, the first processing module comprises:

[0051] An exit lane centerline determination sub-module is configured to: in the predecessor lane centerlines, filter a predecessor lane centerline with a distance to an exit end face of the predecessor road less than a first threshold value; project an end point of the filtered predecessor lane centerline to the learned trajectory as a reference line to obtain a first lateral distance value to the learned trajectory after projection; and determine the predecessor lane centerline with the smallest first lateral distance value as the learned trajectory on the exit lane centerline of the predecessor road.

[0052] The third aspect of the present application provides a vehicle, comprising:

[0053] A processor; and

[0054] A memory having executable code stored thereon, which, when executed by the processor, causes the processor to perform the method described above.

[0055] The fourth aspect of the present application provides a computer-readable storage medium having executable code stored thereon, which, when executed by a processor of a vehicle, causes the processor to perform the method described above.

[0056] The fifth aspect of the present application provides a computer program product comprising computer instructions, which, when executed by a processor, implement the method described above.

[0057] The technical solution provided by the present application can have the following beneficial effects:

[0058] The technical scheme of the present application comprises the following steps: acquiring a constructed road, acquiring a lane center line in the road, and acquiring a learning trajectory; then determining the learning trajectory on an exit lane center line of the preceding road and determining the learning trajectory on an entrance lane center line of the subsequent road; then generating a topological connection line between the preceding road and the subsequent road according to the exit lane center line and the entrance lane center line and referring to the learning trajectory; and finally obtaining a road network topology between the preceding road and the subsequent road according to the topological connection line. The present application constructs a road network topology by using a learning trajectory, learns from prior information of the learning trajectory, and constructs a topology and generates a topological connection line between roads by determining an exit lane center line of a preceding road and an entrance lane center line of a subsequent road, so that the construction quality of the road network topology can be improved, and the safety and experience of autonomous driving can be improved.

[0059] It should be understood that the foregoing general description and the following detailed description are only exemplary and explanatory, and are not limiting to the present application. BRIEF DESCRIPTION OF DRAWINGS

[0060] The above and other objects, features and advantages of the present application will become more apparent from the following detailed description when taken in conjunction with the accompanying drawings in which like reference characters refer to like parts throughout and in which:

[0061] Fig. 1 is a flowchart of a method for constructing a road network topology according to an embodiment of the present application;

[0062] Fig. 2 is a flowchart of a method for constructing a road network topology according to another embodiment of the present application;

[0063] Fig. 3 is a first schematic diagram of a method for constructing a road network topology according to an embodiment of the present application;

[0064] Fig. 4 is a second schematic diagram of a method for constructing a road network topology according to an embodiment of the present application;

[0065] Fig. 5 is a third schematic diagram of a method for constructing a road network topology according to an embodiment of the present application;

[0066] Fig. 6 is a fourth schematic diagram of a method for constructing a road network topology according to an embodiment of the present application;

[0067] Fig. 7 is a fifth schematic diagram of a method for constructing a road network topology according to an embodiment of the present application;

[0068] Fig. 8 is a sixth schematic diagram of a method for constructing a road network topology according to an embodiment of the present application;

[0069] Fig. 9 is a seventh schematic diagram of a method for constructing a road network topology according to an embodiment of the present application;

[0070] FIG. 10 is a structural schematic diagram of a road network topology construction device according to an embodiment of the present application;

[0071] FIG. 11 is a structural schematic diagram of a road network topology construction device according to another embodiment of the present application;

[0072] FIG. 12 is a structural schematic diagram of a vehicle according to an embodiment of the present application. DETAILED DESCRIPTION

[0073] Embodiments of the present application will be described in detail with reference to the accompanying drawings. Although embodiments of the present application are shown in the drawings, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments described herein. Rather, these embodiments are provided so that the present application is more thoroughly and completely conveyed to those skilled in the art.

[0074] The terms used in the present application are merely for the purpose of describing specific embodiments and are not intended to limit the present application. The singular forms "a," "an," and "the" used in the present application and the appended claims are also intended to include the plural forms, unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used herein refers to and includes any or all possible combinations of one or more associated listed items.

[0075] It should be understood that although the terms "first," "second," "third," etc. can be used in this application to describe various information, these information should not be limited by these terms. These terms are only used to distinguish one type of information from another. For example, the first information can also be referred to as the second information, and similarly, the second information can also be referred to as the first information without departing from the scope of the present application. Therefore, the features defined with "first," "second" can explicitly or implicitly include one or more of the features. In the description of the present application, the meaning of "a plurality of" is two or more, unless otherwise specifically limited.

[0076] Road network topology is crucial for route planning and navigation in autonomous driving. The construction quality of road network topology can affect the safety and experience of autonomous driving. How to improve the construction quality of road network topology is a problem to be solved in autonomous driving technology.

[0077] The present application provides a road network topology construction method, which can improve the construction quality of road network topology and improve the safety and experience of autonomous driving.

[0078] The technical solutions of the embodiments of the present application will be described in detail below with reference to the accompanying drawings.

[0079] FIG. 1 is a flowchart of a method for constructing a road network topology according to an embodiment of the present application. The embodiment of the present application constructs a road network topology between roads based on a reference learning trajectory, on the premise that the roads and lanes have been constructed.

[0080] Referring to FIG. 1, the method includes:

[0081] S101, obtaining constructed roads and lane centerlines in the roads, wherein the roads include a predecessor road and a successor road.

[0082] The roads can be constructed by a preset module, and can be constructed based on information detected by perception and using related technologies. The embodiment of the present application does not limit the road construction method.

[0083] The roads include lanes, and the lanes have lane centerlines. Generally, a road including several lanes corresponds to several road centerlines.

[0084] Generally, two roads are needed to construct a network topology, one as a predecessor and one as a successor. The first road appearing in front of the ego vehicle can be defined as the predecessor road, and the road following the predecessor road can be referred to as the successor road. The predecessor road includes a predecessor lane centerline, and the successor road includes a successor lane centerline.

[0085] S102, obtaining a learning trajectory.

[0086] The technical solution of the embodiment of the present application constructs a network topology between roads by taking a learning trajectory as a reference. The learning trajectory in the embodiment of the present application can be a driving trajectory of a vehicle in a commuting mode, i.e., an AI chauffeur.

[0087] The embodiment of the present application can obtain a learning trajectory of a user. The learning trajectory of the user can be obtained at the vehicle end or uploaded from the cloud end. The learning trajectory can include recorded trajectory information of a vehicle driven by the ego vehicle, wherein the trajectory information includes coordinate information of a series of trajectory points sorted according to time, and the trajectory points are all recorded trajectory points in a learning route.

[0088] S103, determining an exit lane centerline of the predecessor road in the learning trajectory and determining an entry lane centerline of the successor road in the learning trajectory.

[0089] The exit lane centerline of the predecessor road in the learning trajectory can be determined by screening, from the predecessor lane centerline, a predecessor lane centerline having a distance to an exit end face of the predecessor road less than a first threshold value.

[0090] The end point of the screened predecessor lane centerline is projected onto the learning trajectory as a reference line to obtain a first lateral distance value from the learning trajectory after the projection.

[0091] The predecessor lane centerline with the minimum first lateral distance value is taken as the learning trajectory at the exit lane centerline of the predecessor road.

[0092] The distance between the successor lane centerline and the entry end surface of the successor road is less than a second threshold value.

[0093] The starting point of the selected successor lane centerline is projected to the learning trajectory to obtain a second lateral distance value after projection from the learning trajectory.

[0094] The successor lane centerline with the minimum second lateral distance value is taken as the learning trajectory at the entry lane centerline of the successor road.

[0095] S104, according to the exit lane centerline and the entry lane centerline, and referring to the learning trajectory, a topological connection line of the predecessor road and the successor road is generated.

[0096] The exit lane centerline and the entry lane centerline are constructed into a first topological relationship, and a first topological connection line of the predecessor road and the successor road is generated by referring to the learning trajectory.

[0097] The lane centerlines of the predecessor road except the exit lane centerline are constructed into a second topological relationship with the lane centerlines of the successor road except the entry lane centerline, and a second topological connection line of the predecessor road and the successor road is generated by referring to the learning trajectory.

[0098] The lane centerlines of the predecessor road except the exit lane centerline are constructed into a second topological relationship with the lane centerlines of the successor road except the entry lane centerline, including:

[0099] The lane centerlines of the predecessor road except the exit lane centerline are projected to the learning trajectory to obtain a third lateral distance value after projection from the learning trajectory.

[0100] The lane centerlines of the successor road except the entry lane centerline are projected to the learning trajectory to obtain a fourth lateral distance value after projection from the learning trajectory.

[0101] The two lane centerlines corresponding to the minimum deviation between the third lateral distance value and the fourth lateral distance value are constructed into a second topological relationship.

[0102] According to a preset smoothing algorithm, a smoothing trajectory is generated as a topological connection line of the predecessor road and the successor road from the end point of the exit lane centerline to the starting point of the entry lane centerline, with the learning trajectory as a reference.

[0103] S105, according to the topological connection line, a road network topology between the predecessor road and the successor road is obtained.

[0104] The road network topology between the predecessor road and the successor road can be obtained according to the first topology connection line and the second topology connection line.

[0105] The technical scheme of the embodiment of the present application obtains the constructed road and the lane center line in the road, and obtains the learning trajectory. Then, the exit lane center line of the learning trajectory in the predecessor road is determined, and the entry lane center line of the learning trajectory in the successor road is determined. Then, the topology connection line of the predecessor road and the successor road is generated according to the exit lane center line and the entry lane center line and referring to the learning trajectory. Finally, the road network topology between the predecessor road and the successor road is obtained according to the topology connection line. The road network topology is constructed by using the learning trajectory, the prior information of the learning trajectory is used for reference, and the topology is constructed and the topology connection line between the roads is generated by determining the exit lane center line of the predecessor road and the entry lane center line of the successor road. Therefore, the construction quality of the road network topology can be improved, and the safety of the autonomous driving and the experience of the autonomous driving can be improved.

[0106] FIG. 2 is a flowchart of a method for constructing a road network topology according to another embodiment of the present application.

[0107] The technical scheme of the embodiment of the present application constructs the road network topology between the roads by using the learning trajectory on the basis of the road (road) and the lane (lane center line) that have been constructed based on the environmental perception information. The network topology between the roads is constructed based on the learning trajectory. According to the network topology and the topology connection line, the path preferred by the driver can be generated for the user to use, and the generated path can be safe, passable and in line with the driving preference of the driver.

[0108] Referring to FIG. 2, the method comprises the following steps.

[0109] S201, obtaining a constructed road and a lane center line in the road, wherein the road comprises a predecessor road and a successor road.

[0110] Referring to FIG. 3, the two rectangular frames in FIG. 3 are the road (road) that has been constructed, and the dashed lines in the rectangular frames are the lanes (lane center lines) in the constructed road. The road is the entire drivable area, and the road contains multiple lanes. The lane is the center line of the lane, that is, the lane center line. Therefore, a road contains several lanes, which corresponds to several lanes.

[0111] For the two roads in FIG. 3, for the convenience of description, the embodiments of the present application define them as a parent road (a predecessor road, i.e., the first road in front of the vehicle in the figure) and a child road (a successor road, i.e., the second road in front of the vehicle in the figure). That is, the first road appearing in front of the vehicle is defined as the parent road, and the road following the parent road is called the child road, and the order is generally irreversible.

[0112] Generally, two roads are needed to build a network topology, one being a predecessor and the other being a successor. That is, if the blank area between the two roads is regarded as a virtual connecting road, the parent road can be understood as a predecessor road, and the child road can be understood as a successor road.

[0113] The embodiments of the present application use the different names of parent and child, mainly to reflect the relationship of the topology network. Generally, the construction of the network is to find the corresponding relationship from all the lanes of the predecessor road to all the lanes of the successor road.

[0114] The road corresponding to the rectangular frame that has been constructed can be constructed by a preset module, can be constructed based on the information detected by perception and by using related technologies, and the road construction method is not limited by the embodiments of the present application. For example, in related technologies, the cloud and the vehicle end can independently construct the road structure, wherein the cloud can generate a cloud road structure relying on the advantage of over-the-horizon, and the vehicle end can generate a vehicle end road structure relying on the perception information of the vehicle end. Alternatively, the cloud and the vehicle end cooperate to construct the road structure.

[0115] Among them, the lane in the parent road can be called a parent lane, and the lane in the child road can be called a child lane, that is, the predecessor road includes a predecessor lane center line, and the successor road includes a successor lane center line.

[0116] S202, acquiring a learning trajectory.

[0117] In order to ensure the connectivity of the road, the network topology structure of the road generally needs to be constructed. In order to ensure that at least one passable path is provided according to the network topology structure, the embodiments of the present application construct the network topology between roads by taking the learning trajectory as a reference.

[0118] The learning trajectory of the vehicle of the user can be acquired in the embodiments of the present application. The learning trajectory of the user can be acquired at the vehicle end or uploaded from the cloud end. The learning trajectory can include trajectory information recorded when the vehicle travels, and the trajectory information includes coordinate information of a series of trajectory points sorted according to time.

[0119] The learning trajectory in the embodiments of the present application can be a driving trajectory in a commuting mode of the vehicle, that is, an AI chauffeur. The commuting mode in the embodiments of the present application can be understood in a broad sense as a driving mode adopted by the user on a commuting route on weekdays and a driving mode adopted by the user on a high-frequency driving route between certain common places on weekends or holidays, and is not limited to travel on weekdays. For example, the user frequently goes back and forth between a residence and a fixed entertainment place on weekends, and the driving trajectory with the residence and the entertainment place as start and end points is also applicable to the commuting mode. By using the AI chauffeur function, the learning trajectory can be recorded and generated through one-time learning, and a certain number of learning trajectories can be stored, for example, 5 or 10 learning trajectories can be stored at the vehicle end or the cloud end. The route length of each learning trajectory can be a preset length, for example, can be up to 100 kilometers but is not limited thereto.

[0120] In S203, the exit lane center line of the learning trajectory on the predecessor road and the entry lane center line of the learning trajectory on the successor road are determined.

[0121] The embodiments of the present application can find the exit lane of the learning trajectory on the predecessor road and the entry lane of the learning trajectory on the successor road according to the physical relationship between each lane and the learning trajectory. Referring to FIG. 4, the red line 401 represents the user trajectory, that is, the learning trajectory.

[0122] The learning trajectory can be used as one of the external inputs for constructing the network topology of the road, and the lane is used as an internal input for constructing the network topology. There are many lanes in the road, but in the embodiments of the present application, the construction of the network topology mainly concerns the lane on the end face of the road.

[0123] In the predecessor lane center line, the predecessor lane center line with a distance from the exit end face of the predecessor road less than a first threshold value can be filtered out; and in the successor lane center line, the successor lane center line with a distance from the entry end face of the successor road less than a second threshold value can be filtered out. The first threshold value and the second threshold value can be set according to experience or actual needs.

[0124] For example, referring to FIG. 5, a parent lane that is very close to the exit end surface of the parent road can be screened out according to the distance. Similarly, a child lane that is very close to the entrance end surface of the child road can be screened out. The screened-out parent lane and child lane are used as internal input for constructing the network topology.

[0125] The lane can be on the left or right of the learning trajectory, and can be far or close to the learning trajectory. In the embodiment of the present application, the exit lane and the entrance lane of the learning trajectory are searched first.

[0126] The end point of the screened-out predecessor lane center line can be projected to the learning trajectory to obtain a first lateral distance value from the learning trajectory, taking the learning trajectory as a reference line.

[0127] The predecessor lane center line with the smallest first lateral distance value is taken as the exit lane center line of the learning trajectory on the predecessor road.

[0128] The start point of the screened-out successor lane center line can be projected to the learning trajectory to obtain a second lateral distance value from the learning trajectory, taking the learning trajectory as a reference line.

[0129] The successor lane center line with the smallest second lateral distance value is taken as the entrance lane center line of the learning trajectory on the successor road.

[0130] For example, the learning trajectory can be taken as a reference line, and the last point of the previously screened-out parent lane can be projected to the trajectory. Each parent lane will output a lateral distance value after projection. Similarly, the first point of the previously screened-out child lane can be projected to the trajectory, and each child lane will also output a lateral distance value after projection.

[0131] Then, the lane with the smallest lateral distance value in the parent lane is selected as the exit lane of the learning trajectory, and the lane with the smallest lateral distance value in the child lane is selected as the entrance lane of the learning trajectory.

[0132] As shown in FIG. 5, the three yellow small arrows in the parent lane point to the last point of the three parent lanes, and the three green small arrows in the child lane point to the first point of the three child lanes. The labels l0, l1, l2 in FIG. 5 are respectively the lateral distance values obtained by projecting the three child lanes onto the trajectory, and the labels l0', l1', l2' in FIG. 5 are respectively the lateral distance values obtained by projecting the three parent lanes onto the trajectory.

[0133] From the comparison result after FIG. 5, it can be known that the exit lane of the learning trajectory is the leftmost lane in the parent road with the lateral distance value l0', and the entry lane of the learning trajectory is the left lane in the child road with the lateral distance value l0.

[0134] S204, constructing a first topological relationship between the exit lane center line and the entry lane center line, and generating a first topological connection line between the predecessor road and the successor road by referring to the learning trajectory.

[0135] In the method, the first topological relationship can be constructed between the exit lane center line and the entry lane center line, and the first topological connection line between the predecessor road and the successor road can be generated by referring to the learning trajectory.

[0136] After obtaining the exit lane of the learning trajectory and the entry lane of the learning trajectory, the exit lane and the entry lane are taken as a group of topologies, a first topological relationship is constructed, and a first topological connection line between roads is constructed between the two lanes by referring to the shape of the learning trajectory. See FIG. 6.

[0137] In FIG. 6, the black arrow connection line 601 is a topological connection line between two roads, which is equivalent to a path, that is, a vehicle can drive from the leftmost lane of the parent road to the leftmost lane of the child road according to the path formed by the topological connection line, and the topological selection is the most reasonable in the current straight driving condition, and no invalid lane changing is needed.

[0138] S205, constructing a second topological relationship between the lane center lines of the predecessor road except the exit lane center line and the lane center lines of the successor road except the entry lane center line, and generating a second topological connection line between the predecessor road and the successor road by referring to the learning trajectory.

[0139] In the method, the second topological relationship can be constructed between the lane center lines of the predecessor road except the exit lane center line and the lane center lines of the successor road except the entry lane center line, and the second topological connection line between the predecessor road and the successor road can be generated by referring to the learning trajectory.

[0140] For example, the lane center line other than the exit lane center line of the preceding road is projected to the learning trajectory to obtain a third lateral distance value after projection from the learning trajectory;

[0141] The lane center line other than the entrance lane center line of the subsequent road is projected to the learning trajectory to obtain a fourth lateral distance value after projection from the learning trajectory;

[0142] The two lane center lines corresponding to the minimum deviation between the third lateral distance value and the fourth lateral distance value are used to construct a second topological relationship.

[0143] The embodiments of the present application construct a first topological relationship by taking the exit lane and the entrance lane as a group of topologies, and after the first topological connection line between the exit lane and the entrance lane is constructed with reference to the shape of the learning trajectory, the remaining lanes can continue to be taken as a new group of topologies based on the learning trajectory, and the lanes having a distance substantially consistent with the learning trajectory in the physical position are taken as the next group of topologies.

[0144] At this time, the processing method can be consistent with the method of finding the exit lane and the entrance lane. For example, the remaining lanes can also be projected to the learning trajectory to obtain respective lateral distance values, and then the two lanes corresponding to the minimum deviation of the lateral distance values are found as the input of a new group of topologies.

[0145] For example, in FIG. 5, it can be clearly seen that the difference between the lateral distance values l1 and l1' is significantly smaller than the difference between the lateral distance values l1 and l2', and smaller than the difference between the lateral distance values l1 and l0'. Therefore, the two lanes corresponding to the lateral distance values l1 and l1' are a new group of network topologies. Other lanes are similar, for example, the two lanes corresponding to the lateral distance values l2 and l2' are a new group of network topologies.

[0146] Further referring to FIG. 7, from the position relationship, it can be known that the two lanes with the same color of the two arrows can constitute two groups of topologies, that is, the two lanes marked as 71 and 71' corresponding to the yellow arrows constitute a group of topologies, and the two lanes marked as 72 and 72' corresponding to the blue arrows constitute a group of topologies.

[0147] Among them, the group of topologies constructed by the two lanes marked as 71 and 71' can construct a second topological connection line between the two lanes with reference to the shape of the learning trajectory; the group of topologies constructed by the two lanes marked as 72 and 72' can also construct a second topological connection line between the two lanes with reference to the shape of the learning trajectory.

[0148] Wherein, the embodiment of the present application can generate a smooth trajectory as the topological connection line of the predecessor road and the successor road from the end point of the exit lane center line to the starting point of the entrance lane center line, taking the learning trajectory as the reference according to the preset smoothing algorithm.

[0149] Wherein, the preset smoothing algorithm may be, for example, a smoothing algorithm based on curve interpolation (polynomial curve algorithm, spline curve algorithm), a sampling-based A-star algorithm, etc.

[0150] For example, the A-star algorithm is a commonly used heuristic search algorithm, which has good efficiency and accuracy, and is widely used in the field of path planning. The inflection point in the path information generated by the A-star algorithm can be processed by arc to realize the smoothing of the path.

[0151] It should be noted that the above scheme description is an example described in the simplest straight line case and the number of parent lanes and the number of child lanes is equal. In reality, various more complex working conditions may be encountered, and the technical scheme of the embodiment of the present application is based on the construction of road network topology from user trajectory, so that in many cases, the optimal solution can be provided.

[0152] For example, turning around is a common working condition, but without the support of high-precision maps, it is difficult to give a suitable reference path only by perception information, but by using the technical scheme of the present application, the reference user trajectory, i.e. learning trajectory, is processed to construct the network topology between roads, which can provide great help for generating path.

[0153] Taking FIG. 8 as an example, the blue dashed line 801 with arrow is the topological connection line of the constructed network topology, and the red dashed line 802 is the user trajectory, i.e. learning trajectory, which is the driving trajectory in the turning scenario. By using the technical scheme of the present application, the exit lane and the entrance lane of the learning trajectory can be found, and then the shape of the learning trajectory can be referred to, so as to construct a topological connection line almost the same as the learning trajectory. The topological connection line is equivalent to a generated reference path, so the turning scenario can also ensure that the vehicle passes through the turning area very smoothly and comfortably.

[0154] Because the learning trajectory has been obtained, between the exit lane and the entrance lane, the embodiment of the present application can utilize the designed path smoothing algorithm, taking the learning trajectory as the reference, referring to the shape of the learning trajectory, from the last point of the parent lane to the first point of the child lane, to generate a point similar to the shape of the learning trajectory as the final output topological connection line, which is the final output topological path.

[0155] It should be noted that in order to show the effect in FIG. 8, the topological path (planned path) and the learning trajectory given in FIG. 8 are provided with a point gap, but in fact the topological path (planned path) and the learning trajectory are basically consistent. Similarly, the way of constructing the road network topology with reference to the learning trajectory in the embodiments of the present application is also similar to the processing of other scenes such as some main and auxiliary road switching scenes and complex scenes such as left turn and right turn.

[0156] S206, obtaining the road network topology between the predecessor road and the successor road according to the first topological connection line and the second topological connection line.

[0157] Among them, according to the first topological connection line and the second topological connection line, the complete road network topology between the predecessor road and the successor road can be obtained.

[0158] The first topological relationship and the first topological connection line constructed according to the exit lane and the entrance lane, and the second topological relationship and the second topological connection line constructed according to the other lanes except the exit lane and the entrance lane, can obtain the complete network topology structure between two roads. The information contained in the complete network topology structure between two roads generally includes information such as how many parent lanes can go out, how many child lanes can go in, and which child lane is connected to after each parent lane goes out.

[0159] The complete network topology structure between two roads can be seen in FIG. 9. In FIG. 9, the blue lines with arrows 91, 92 and 93 are the topological connection lines of the constructed network topology, and the red trajectory 901 is the user trajectory, i.e. the learning trajectory.

[0160] It should be noted that after obtaining the complete road network topology between the predecessor road and the successor road, the topological connection line in the network topology can be used as a driving path between roads for reference selection during automatic driving, or a driving path can be generated again according to the topological connection line for reference selection during automatic driving, so that the user is given multiple path options while at least one learning trajectory is followed during automatic driving, thereby improving the flexibility and traffic capacity of automatic driving.

[0161] It can be found that the embodiment of the application constructs the road network topology based on the learning trajectory, and finally outputs the topological path, which can not only ensure the safety of autonomous driving, but also meet the kinematics and dynamics constraints of the vehicle. Because the learning trajectory is a driving trajectory learned by the user and uploaded as a learning trajectory after being approved by the user, the learning trajectory can be confirmed as a safe and available route that meets the user's preferences from the user's perspective. Therefore, the topological path processed according to the user's learning results, i.e., the learning trajectory, can also ensure that the vehicle can safely drive through, for example, the U-turn scene can also ensure that the vehicle can pass through the U-turn area very smoothly and comfortably.

[0162] The embodiment of the application constructs the road network topology based on the learning trajectory, which can provide stable and reliable input for path planning and prediction, so that the vehicle can follow the learning trajectory while having other more path options, thereby improving the flexibility and traffic capacity of autonomous driving. The embodiment of the application provides multiple options for the user while at least one learning trajectory can be followed, which not only ensures that there is a path to walk, but also increases the diversity of choices and improves the user's autonomous driving experience.

[0163] Corresponding to the foregoing application function implementation method embodiment, the application also provides a road network topology construction device, a vehicle and corresponding embodiments.

[0164] FIG. 10 is a structural schematic diagram of a road network topology construction device according to an embodiment of the application.

[0165] Referring to FIG. 10, the road network topology construction device 100 provided by the embodiment of the application includes a first acquisition module 101, a second acquisition module 102, a first processing module 103, a second processing module 104, and a result generation module 105.

[0166] The first acquisition module 101 is configured to acquire a constructed road and acquire a lane center line in the road, wherein the road includes a precedent road and a successor road. The road can be constructed by a preset module, and can be constructed based on information detected by perception and using related technologies, and the road construction method is not limited in the embodiment of the application. The road includes a lane, and the lane has a lane center line. Generally, a road including several lanes corresponds to several road center lines. The first road appearing in front of the ego vehicle can be defined as the precedent road, and the road after the precedent road can be referred to as the successor road. The precedent road includes a precedent lane center line, and the successor road includes a successor lane center line.

[0167] The second acquisition module 102 is configured to acquire a learning trajectory. The technical solution of the embodiment of the application constructs the network topology between roads by taking the learning trajectory as a reference. The learning trajectory in the embodiment of the application can be a driving trajectory under the vehicle commuting mode, i.e., AI driving.

[0168] The first processing module 103 is configured to determine an exit lane center line of the learning trajectory on the predecessor road and determine an entrance lane center line of the learning trajectory on the successor road.

[0169] The second processing module 104 is configured to generate a topological connection line of the predecessor road and the successor road according to the exit lane center line and the entrance lane center line and in reference to the learning trajectory. The second processing module 104 can construct a first topological relationship between the exit lane center line and the entrance lane center line, generate a first topological connection line of the predecessor road and the successor road in reference to the learning trajectory, and construct a second topological relationship between lane center lines of the predecessor road except the exit lane center line and lane center lines of the successor road except the entrance lane center line, and generate a second topological connection line of the predecessor road and the successor road in reference to the learning trajectory.

[0170] The result generation module 105 is configured to obtain a road network topology between the predecessor road and the successor road according to the topological connection line. The result generation module 105 can obtain a road network topology between the predecessor road and the successor road according to the first topological connection line and the second topological connection line.

[0171] The device provided in the application obtains a constructed road, obtains a lane center line in the road, and obtains a learning trajectory. Then, an exit lane center line of the learning trajectory on a predecessor road is determined, and an entrance lane center line of the learning trajectory on a successor road is determined. Then, a topological connection line of the predecessor road and the successor road is generated according to the exit lane center line and the entrance lane center line and in reference to the learning trajectory. Finally, a road network topology between the predecessor road and the successor road is obtained according to the topological connection line. The road network topology is constructed by using the learning trajectory, the prior information of the learning trajectory is used for reference, and the topologies are constructed and the topological connection lines between roads are generated by determining the exit lane center line of the predecessor road and the entrance lane center line of the successor road. Therefore, the construction quality of the road network topology can be improved, and the safety and experience of autonomous driving can be improved.

[0172] FIG. 11 is a structural schematic diagram of a road network topology construction device according to another embodiment of the application.

[0173] Referring to FIG. 11, a road network topology construction device 100 includes a first obtaining module 101, a second obtaining module 102, a first processing module 103, a second processing module 104, and a result generation module 105.

[0174] The second processing module 104 can include a first topological processing submodule 1041 and a second topological processing submodule 1042.

[0175] The first topology processing submodule 1041 is configured to construct a first topology relationship between the exit lane center line and the entrance lane center line, and generate a first topology connection line between the preceding road and the subsequent road according to the learning trajectory.

[0176] The second topology processing submodule 1042 is configured to construct a second topology relationship between the lane center line of the preceding road except the exit lane center line and the lane center line of the subsequent road except the entrance lane center line, and generate a second topology connection line between the preceding road and the subsequent road according to the learning trajectory.

[0177] The result generation module 105 obtains a road network topology between the preceding road and the subsequent road according to the first topology connection line and the second topology connection line.

[0178] The first processing module 103 can include an exit lane center line determination submodule 1031.

[0179] The exit lane center line determination submodule 1031 is configured to filter, from the preceding lane center lines, a preceding lane center line with a distance to an exit end face of the preceding road less than a first threshold value, project an end point of the filtered preceding lane center line to the learning trajectory as a reference line to obtain a first lateral distance value to the learning trajectory after the projection, and determine the preceding lane center line with the smallest first lateral distance value as the exit lane center line of the preceding road.

[0180] The first processing module 103 can include an entrance lane center line determination submodule 1032.

[0181] The entrance lane center line determination submodule 1032 is configured to filter, from the subsequent lane center lines, a subsequent lane center line with a distance to an entrance end face of the subsequent road less than a second threshold value, project a starting point of the filtered subsequent lane center line to the learning trajectory as a reference line to obtain a second lateral distance value to the learning trajectory after the projection, and determine the subsequent lane center line with the smallest second lateral distance value as the entrance lane center line of the subsequent road.

[0182] The device shown in the embodiments of the present application can construct a road network topology based on a learning trajectory, provide stable and reliable input for path planning and prediction, and enable a vehicle to follow the learning trajectory completely while having more path options, thereby improving the flexibility and traffic capacity of automatic driving. Because the learning trajectory is a driving trajectory learned by a user and uploaded as a learning trajectory after being approved by the user, the learning trajectory can be confirmed as a safe and usable path that meets the user's preferences from the user's perspective.

[0183] With regard to the apparatus in the above-described embodiments, a specific manner in which each module performs an operation has been described in detail in the embodiments related to the method, and thus will not be described in detail here.

[0184] FIG. 12 is a structural schematic diagram of a vehicle according to an embodiment of the present application.

[0185] Referring to FIG. 12, the vehicle 1000 includes a memory 1010 and a processor 1020.

[0186] The processor 1020 can be a central processing unit (CPU), and can also be other general-purpose processors, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor.

[0187] The memory 1010 can include various types of storage units such as a system memory, a read-only memory (ROM), and a permanent storage device. Among them, the ROM can store static data or instructions required by the processor 1020 or other modules of the computer. The permanent storage device can be a rewritable storage device. The permanent storage device can be a non-volatile storage device that does not lose stored instructions and data even after the computer is powered off. In some embodiments, the permanent storage device employs a mass storage device (e.g., a magnetic or optical disk, a flash memory) as a permanent storage device. In some other embodiments, the permanent storage device can be a removable storage device (e.g., a floppy disk, an optical drive). The system memory can be a readable and writable storage device or a volatile readable and writable storage device such as a dynamic random access memory. The system memory can store some or all of the instructions and data required by the processor during runtime. In addition, the memory 1010 can include a combination of any computer readable storage media, including various types of semiconductor storage chips (e.g., DRAM, SRAM, SDRAM, flash memory, programmable read-only memory), magnetic disks and / or optical disks. In some embodiments, the memory 1010 can include a readable and / or writable removable storage device such as a compact disc (CD), a read-only digital versatile disc (e.g., DVD-ROM, double-layer DVD-ROM), a read-only Blu-ray disc, an ultra-density optical disc, a flash memory card (e.g., an SD card, a min SD card, a Micro-SD card, etc.), a magnetic floppy disk, etc. The computer readable storage medium does not include a carrier wave and an instantaneous electronic signal transmitted by wireless or wired transmission.

[0188] The memory 1010 stores executable code, which, when processed by the processor 1020, can cause the processor 1020 to perform part or all of the above-mentioned methods.

[0189] In addition, the method according to the present application can also be implemented as a computer program or a computer program product, which includes computer program code instructions for executing part or all of the steps of the above-mentioned methods of the present application.

[0190] Alternatively, the present application can also be implemented as a computer readable storage medium (or non-transitory machine readable storage medium or machine readable storage medium) having executable code (or computer program or computer instruction code) stored thereon, which, when executed by a processor of a vehicle (or electronic equipment, or a server, etc.), causes the processor to perform part or all of the steps of the above-mentioned methods according to the present application.

[0191] Having described various embodiments of the application, it is to be understood that the above description is meant not to limit and not to encompass all of the possible embodiments. Many modifications and variations of this application can be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. It is intended that the scope of the application be defined by the scope of the patent and by the claims as allowed by the patent office, which can include adaptations based on the description, equivalents, and / or substitutions of elements individually or collectively to the entire disclosure.

Claims

1. A method for constructing a road network topology, characterized in that, include: Obtain the constructed road and the lane centerline in the road, wherein the road includes the preceding road and the succeeding road; Obtain the learning trajectory; Determine the centerline of the exit lane of the preceding road and the centerline of the entrance lane of the following road for the learning trajectory. Based on the center line of the exit lane and the center line of the entrance lane, and with reference to the learning trajectory, a topological connection line between the preceding road and the following road is generated. Based on the topological connection lines, the road network topology between the preceding road and the succeeding road is obtained.

2. The method according to claim 1, characterized in that, The step of generating the topological connection line between the preceding road and the succeeding road based on the center line of the exit lane and the center line of the entrance lane, and with reference to the learning trajectory, includes: A first topological relationship is established between the centerline of the exit lane and the centerline of the entrance lane, and a first topological connection line between the preceding road and the following road is generated with reference to the learning trajectory. A second topological relationship is constructed between the center lines of the preceding road (excluding the center line of the exit lane) and the center lines of the subsequent road (excluding the center line of the entrance lane), and a second topological connection line between the preceding road and the subsequent road is generated with reference to the learning trajectory. The step of obtaining the road network topology between the preceding road and the succeeding road based on the topological connection line includes: Based on the first topology connection line and the second topology connection line, the road network topology between the preceding road and the succeeding road is obtained.

3. The method according to claim 1, characterized in that, The preceding road includes a centerline of the preceding lane, and determining the learning trajectory along the centerline of the exit lane of the preceding road includes: Among the leading lane centerlines, those leading lane centerlines whose distance from the exit end face of the leading road is less than a first threshold are selected. Using the learning trajectory as a baseline, the end point of the selected forward lane centerline is projected onto the learning trajectory to obtain the first lateral distance value between the projected point and the learning trajectory. The centerline of the leading lane with the smallest first lateral distance value is used as the learning trajectory for the centerline of the exit lane of the leading road.

4. The method according to claim 1, characterized in that, The subsequent road includes the centerline of the subsequent lane, and determining the learning trajectory at the centerline of the entrance lane of the subsequent road includes: Among the subsequent lane centerlines, those whose distance from the entrance end face of the subsequent road is less than a second threshold are selected. Using the learning trajectory as a baseline, the starting point of the selected subsequent lane centerline is projected onto the learning trajectory to obtain the second lateral distance value between the projected lane and the learning trajectory. The centerline of the subsequent lane with the smallest second lateral distance value is used as the learning trajectory for the centerline of the entrance lane of the subsequent road.

5. The method according to claim 2, characterized in that, The step of constructing a second topological relationship between the centerlines of the preceding road (excluding the exit lane centerline) and the centerlines of the subsequent road (excluding the entrance lane centerline) includes: Project the center lines of the lanes of the preceding road, excluding the center line of the exit lane, onto the learning trajectory to obtain the third lateral distance value between the projected lane and the learning trajectory. Project the center lines of the subsequent roads, excluding the center line of the entrance lane, onto the learning trajectory to obtain the fourth lateral distance value between the projected lane and the learning trajectory. The second topological relationship is constructed by the two lane centerlines corresponding to the minimum deviation between the third lateral distance value and the fourth lateral distance value.

6. The method according to any one of claims 1 to 5, characterized in that, The step of generating the topological connection line between the preceding road and the succeeding road based on the center line of the exit lane and the center line of the entrance lane, and with reference to the learning trajectory, includes: Based on a preset smoothing algorithm and with the learning trajectory as a reference, a smooth trajectory is generated from the end point of the exit lane centerline to the starting point of the entrance lane centerline as a topological connection line between the preceding road and the following road.

7. The method according to any one of claims 1 to 5, characterized in that: The preceding road is the first road that appears in front of the vehicle, and the following road is the road that follows the preceding road.

8. The method according to claim 6, characterized in that, The preset smoothing algorithm includes a smoothing algorithm based on curve interpolation or an A* algorithm based on sampling.

9. The method according to any one of claims 1 to 5, characterized in that, The constructed roads are generated by the cloud or vehicle-side, or by a combination of cloud and vehicle-side construction.

10. The method according to any one of claims 1 to 5, characterized in that, After obtaining the road network topology between the preceding road and the succeeding road based on the topological connection line, the method further includes: The aforementioned topological connecting lines can be used as a reference for selection during autonomous driving; or... The driving path is regenerated based on the topological connection lines for reference selection during autonomous driving.

11. A road network topology construction device, characterized in that, include: The first acquisition module is used to acquire the constructed road and acquire the lane centerline in the road, wherein the road includes the preceding road and the following road; The second acquisition module is used to acquire the learning trajectory; The first processing module is used to determine the center line of the exit lane of the learning trajectory on the preceding road and the center line of the entrance lane of the learning trajectory on the subsequent road. The second processing module is used to generate a topological connection line between the preceding road and the following road based on the center line of the exit lane and the center line of the entrance lane, and with reference to the learning trajectory. The result generation module is used to obtain the road network topology between the preceding road and the succeeding road based on the topology connection line.

12. The apparatus according to claim 11, characterized in that, The second processing module includes: The first topology processing submodule is used to construct a first topological relationship between the center line of the exit lane and the center line of the entrance lane, and generate a first topological connection line between the preceding road and the following road by referring to the learning trajectory. The second topology processing submodule is used to construct a second topology relationship between the lane centerlines of the preceding road (excluding the exit lane centerline) and the lane centerlines of the subsequent road (excluding the entrance lane centerline), and generate a second topology connection line between the preceding road and the subsequent road with reference to the learning trajectory. The result generation module obtains the road network topology between the preceding road and the succeeding road based on the first topology connection line and the second topology connection line.

13. A vehicle, characterized in that, include: processor; as well as A memory having executable code stored thereon, which, when executed by the processor, causes the processor to perform the method as described in any one of claims 1-10.

14. A computer-readable storage medium, characterized in that: It stores executable code that, when executed by the vehicle's processor, causes the processor to perform the method as described in any one of claims 1-10.

15. A computer program product, characterized in that, The computer program product includes computer instructions that, when executed by a processor, implement the method of any one of claims 1-10.

Images