Road model generation method and apparatus, and vehicle and storage medium

Abstract

The present application relates to a road model generation method and apparatus, and a vehicle and a storage medium. The method comprises: acquiring a first road model generated by a cloud side, wherein the first road model at least comprises first road geometric information and first road topological information; acquiring vehicle-sensed map information detected by a vehicle side; matching the vehicle-sensed map information with the first road geometric information and the first road topological information of the first road model, so as to obtain target road geometric information and target road topological information; and on the basis of the target road geometric information and the target road topological information, generating a vehicle-side target road model. The solution provided in the present application improves the accuracy and stability of a road geometric structure on a vehicle side, and also improves the efficiency of target road model construction by utilizing a priori information of a first road model from an a priori cloud side.

Description

Road model generation method and device, automobile and storage medium

[0001] The present application claims priority to the Chinese patent application No. 202410742616X filed on June 7, 2024 with the State Intellectual Property Office, and entitled "Road model generation method and device, automobile and storage medium", the whole content of which is incorporated herein by reference. TECHNICAL FIELD

[0002] The present application relates to the technical field of autonomous driving, and in particular to a road model generation method and device, an automobile and a storage medium. BACKGROUND

[0003] With the continuous development of autonomous driving technology, road models play an increasingly important role in vehicle navigation and control. Compared with ordinary high-definition maps, road models not only contain detailed road data such as road information, traffic signs, lane lines, traffic signal lights, but also include road topology information of user driving decision information.

[0004] In actual application scenarios, the cloud and the vehicle each independently construct a road structure, wherein the cloud generates a cloud road structure relying on the advantage of over-the-horizon, and the vehicle generates a vehicle road structure relying on vehicle perception information. The related technology needs to first match the vehicle road structure to the road structure of the cloud, and then assign the prior information of the cloud to the vehicle lane to generate a vehicle road model for the matched lane. However, due to the great limitations and stability of real-time perception information of the vehicle, the road structure constructed by the vehicle in real time is unstable and fluctuant, which makes it difficult to geometrically match the road structure of the cloud, and the stable and effective matching between the vehicle and the cloud cannot be completed.

[0005] SUMMARY

[0006] To solve or partially solve the problems in the related technology, the present application provides a road model generation method and device, an automobile and a storage medium, which can improve the accuracy and stability of generating road models.

[0007] The first aspect of the present application provides a road model generation method, comprising:

[0008] obtaining a first road model generated by a cloud, the first road model at least comprising first road geometric information and first road topology information;

[0009] obtaining vehicle perception map information detected by a vehicle;

[0010] matching the vehicle perception map information with the first road geometric information and the first road topology information of the first road model to obtain target road geometric information and target road topology information;

[0011] According to the target road geometry information and the target road topology information, a target road model at a vehicle end is generated.

[0012] In an embodiment, the vehicle-sensed map information is matched with the first road geometry information and the first road topology information of the first road model to obtain the target road geometry information and the target road topology information, including:

[0013] The vehicle-sensed map information is matched with the first road geometry information to obtain the target road geometry information.

[0014] According to the first road topology information, the vehicle-sensed map information and the target road geometry information, the target road topology information is constructed.

[0015] In an embodiment, the first road topology information at least includes one or more of lane turning information, predecessor and successor information, lane type and user driving trajectory.

[0016] In an embodiment, the vehicle-sensed map information is matched with the first road geometry information to obtain the target road geometry information, including:

[0017] According to the vehicle-sensed map information, second road geometry information is obtained.

[0018] The first road geometry information and the second road geometry information are geometrically matched, and the target road geometry information is generated according to a matching result.

[0019] In an embodiment, the first road geometry information and the second road geometry information are geometrically matched, including:

[0020] A self-vehicle perception reliable range is determined, and within the self-vehicle perception reliable range, the target road geometry information is determined according to the second road geometry information.

[0021] In a case of exceeding the self-vehicle perception reliable range, the second road geometry information is corrected according to the first road geometry information to determine the target road geometry information.

[0022] In an embodiment, in a case of exceeding the self-vehicle perception reliable range, the second road geometry information is corrected according to the first road geometry information to determine the target road geometry information, including:

[0023] According to a preset geometry information conversion format, a first road center line of the first road geometry information and a second road center line of the second road geometry information are obtained.

[0024] The first road center line and the second road center line exceeding the self-vehicle perception reliable range are weightedly fitted to determine the target road geometry information.

[0025] In an embodiment, the target road topology information is constructed according to the first road topology information, the vehicle map information and the target road geometry information, comprising:

[0026] The first road topology information is verified, and the first road topology information that fails to pass the verification is corrected according to the vehicle map information to obtain second road topology information;

[0027] The vehicle map information and the first road topology information are compared to determine new road topology information;

[0028] The target road topology information is generated according to the new road topology information and the second road topology information.

[0029] In an embodiment, the first road topology information is verified, comprising:

[0030] According to a preset judgment rule, the connection rationality of the predecessor and successor information of the first road topology information is judged; and / or;

[0031] According to the vehicle map information, the accuracy of the lane turning information, the predecessor and successor information and the lane type of the first road topology information is verified.

[0032] In an embodiment, the vehicle map information and the first road topology information are compared to determine the new road topology information, comprising:

[0033] According to the vehicle map information, the new user driving track and the new user driving habit are obtained;

[0034] According to the new user driving track and the new user driving habit, the new road topology information is generated.

[0035] The second aspect of the present application provides a road model generation device, comprising:

[0036] A first acquisition module is configured to acquire a first road model generated by a cloud, wherein the first road model at least comprises first road geometry information and first road topology information;

[0037] A second acquisition module is configured to acquire vehicle map information detected by a vehicle terminal;

[0038] A processing module is configured to match the vehicle map information with the first road geometry information and the first road topology information of the first road model to obtain target road geometry information and target road topology information;

[0039] A determination module is configured to generate a target road model of the vehicle terminal according to the target road geometry information and the target road topology information.

[0040] In an embodiment, the processing module comprises:

[0041] The first processing module is configured to match the vehicle-sensed map information with the first road geometry information to obtain target road geometry information.

[0042] The second processing module is configured to construct target road topology information according to the first road topology information, the vehicle-sensed map information and the target road geometry information.

[0043] In an embodiment, the first processing module acquires second road geometry information according to the vehicle-sensed map information, performs geometric matching on the first road geometry information and the second road geometry information, and generates target road geometry information according to a matching result.

[0044] In an embodiment, the first processing module determines a self-vehicle sensing reliable range, determines target road geometry information according to second road geometry information within the self-vehicle sensing reliable range, and corrects the second road geometry information according to the first road geometry information to determine target road geometry information when the self-vehicle sensing reliable range is exceeded.

[0045] The third aspect of the present application provides an automobile, comprising:

[0046] a processor; and

[0047] a memory having executable code stored thereon, which, when executed by the processor, causes the processor to perform the method described above.

[0048] 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 an electronic device, causes the processor to perform the method described above.

[0049] The technical solution provided by the present application can include the following beneficial effects:

[0050] In the first aspect, the present application acquires a first road model generated by a cloud and vehicle-sensed map information detected by a vehicle end; and on the basis of prior information of the first road model, the vehicle-sensed map information is fused to generate a target road model of the vehicle end. Compared with a traditional generation mode of a road model of the vehicle end, the present application assigns road information with prior information of the cloud to road structure of the vehicle end, saves computing power for constructing a road model of the vehicle end, and improves efficiency of generating a target road model.

[0051] In a second aspect, the first road geometry information and the first road topology information of the cloud are retained, so that the cloud has the advantage of constructing a road model with a long-distance view. The vehicle map information, the first road geometry information and the first road topology information are matched, so that the road structure instability caused by the large fluctuation of the vehicle perception information is reduced, and a more stable target road model is constructed. Meanwhile, the vehicle map information can collect the new driving track of the user in real time, and the new driving track of the user is fused into the target road model, so that the target road model can be updated in real time according to the driving characteristics of the user.

[0052] It should be understood that the foregoing general description and the following detailed description are only exemplary and explanatory and are not restrictive of the application. BRIEF DESCRIPTION OF DRAWINGS

[0053] 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:

[0054] FIG. 1 is a flow diagram of a road model generation method according to an embodiment of the present application;

[0055] FIG. 2 is a schematic diagram of a first road model of the cloud according to an embodiment of the present application;

[0056] FIG. 3 is a schematic diagram of first road geometry information of the first road model according to an embodiment of the present application;

[0057] FIG. 4 is a schematic diagram of vehicle map information according to an embodiment of the present application;

[0058] FIG. 5 is a schematic diagram of a target road model according to an embodiment of the present application;

[0059] FIG. 6 is a schematic diagram of a road model generation device according to an embodiment of the present application;

[0060] FIG. 7 is a schematic diagram of a vehicle according to an embodiment of the present application. DETAILED DESCRIPTION

[0061] Embodiments of the present application will be described in more detail by referring to the 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. On the contrary, these embodiments are provided to make the present application more thorough and complete, and to fully convey the scope of the present application to those skilled in the art.

[0062] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting. As used in this application and the appended claims, the singular forms "a," "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "and / or," as used herein, refers to and encompasses any and all possible combinations of one or more of the associated listed items.

[0063] It should be understood that, although the terms "first," "second," "third," etc. can be used in this application to describe various information, the information should not be limited to these terms. These terms are only used to distinguish one piece of information from another piece of information. 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 application. Therefore, the features defined with "first," "second" can explicitly or implicitly include one or more of the features. In the description of the application, the meaning of "a plurality of" is two or more, unless otherwise specifically limited.

[0064] With the continuous development of automatic driving technology, the road model plays an increasingly important role in vehicle navigation and control. Compared with ordinary high-precision maps, the road model not only contains detailed road data such as road information, traffic signs, lane lines, traffic signal lights, but also includes road topology information of user driving decision information.

[0065] In actual application scenarios, the cloud generates a cloud road model relying on the advantage of over-the-horizon, and the vehicle end generates a vehicle end road geometry relying on the vehicle end perception information. The related technology matches the vehicle end road geometry with the cloud road structure. For the matched lane, the prior information of the cloud is given to the vehicle end road model of the vehicle end lane. However, due to the great limitations and stability of the real-time perception information of the vehicle end, the road structure constructed according to the vehicle end is unstable and fluctuant, and it is difficult to geometrically match the road structure of the cloud, so that stable and effective matching between the vehicle end and the cloud cannot be completed.

[0066] The application provides a road model generation method and device, an automobile and a storage medium, which can improve the accuracy and stability of the generated road model.

[0067] The technical solutions of the embodiments of the application are described in detail below with reference to the drawings.

[0068] FIG. 1 is a flow diagram of a road model generation method according to an embodiment of the application.

[0069] Referring to FIG. 1, the road model generation method comprises:

[0070] Step S101: Obtain the first road model generated in the cloud. The first road model includes at least: first road geometric information and first road topology information.

[0071] In this application, the road model in the cloud is a global map constructed by a cloud service provider. It contains information such as roads, buildings, landmarks, traffic flow, and speed limits. The map information of the road model in the cloud is typically used for navigation, route planning, and global positioning. Autonomous driving systems can obtain global information from cloud maps, such as road topology, traffic lights, and zebra crossings.

[0072] In this embodiment, the road model includes geometric information and topological information. Geometric information characterizes the geometric structure of lanes; topological information includes both the physical attributes of the user's driving road and semantic information representing the user's driving preferences. The user's driving semantics can be determined based on the topological information. The physical attributes of the user's driving road include at least one or more of the following: lane turning information, lane type, and lane connectivity. Lane turning information represents lane turning information, such as straight, left turn, right turn, etc. Lane types include at least: bus lane, non-motorized vehicle lane, tidal flow lane, roundabout lane, etc. Lane connectivity uses successor / follower information identifiers to indicate which lane can be used to reach another lane without changing lanes. The semantic information of user driving preferences includes lane priority, lanes the user has passed through during route learning, and low-priority lanes such as the far right or those with poor quality. The above topological information can be stored in the corresponding lane line category using corresponding identifiers.

[0073] In one embodiment, the first road topology information includes at least one or more of the following: lane turning information, preceding and following information, lane type, and user driving trajectory.

[0074] In this embodiment, the successive relationships of lanes are represented, but are not limited to, using an adjacency list stored in a graph, and lane turning information is represented using a combination of symbol types, for example:

[0075] enumLaneTurnType{

[0076] TURN_TYPE_UNKNOWN = 0; (indicating the vehicle's direction is unknown);

[0077] STRAIGHT = 1; (Indicates that the vehicle is going straight).

[0078] TURN_LEFT = 2; (indicates the vehicle is turning left);

[0079] TURN_RIGHT = 3; (indicates that the vehicle is turning right);

[0080] U_TURN = 4; (indicates the vehicle is turning around);

[0081] STRAIGHT_AND_TURN_LEFT = 5; (representing the vehicle straight and then left turn);

[0082] STRAIGHT_AND_TURN_RIGHT = 6; (representing the vehicle straight and then right turn);

[0083] STRAIGHT_AND_UTURN = 7; (representing the vehicle straight and then U-turn);

[0084] STRAIGHT_AND_TURN_LEFT_AND_TURN_RIGHT = 8; (representing the vehicle straight and then left turn and right turn);

[0085] STRAIGHT_AND_TURN_LEFT_AND_UTURN = 9; (representing the vehicle straight and then left turn and U-turn);

[0086] STRAIGHT_AND_TURN_RIGHT_AND_UTURN = 10; (representing the vehicle straight and then right turn and U-turn);

[0087] TURN_LEFT_AND_TURN_RIGHT = 11; (representing the vehicle left turn and then right turn);

[0088] TURN_LEFT_AND_UTURN = 12; (representing the vehicle left turn and then U-turn);

[0089] TURN_RIGHT_AND_UTURN = 13; (representing the vehicle right turn and then left turn).

[0090] The embodiment assigns different lane turning combinations to obtain the turning identification according to the lane, so as to facilitate the rapid reading of the turning information of the lane according to the turning identification from the topological information. It should be noted that the geometric information and the topological information in the present application are composed of a unified data structure, or can be stored in different sets, and the corresponding road information can be found by querying the corresponding set.

[0091] FIG. 2 is a schematic diagram of a cloud-generated road model according to an embodiment of the present application, which includes lanes 1-7, connecting lines 1-4, center lines 1-7, and landmark arrows representing turning information. Each lane corresponds to a center line including lane information, lane n corresponds to center line n, for example, lane 1 corresponds to center line 1, and each center line saves corresponding geometric feature information in the form of a chain. The connecting line is used to represent the corresponding predecessor and successor information, and the direction of the connecting line is used to represent the connection track, for example, connecting line 1 represents the predecessor and successor information from lane 5 to lane 1, connecting line 2 represents the predecessor and successor information from lane 5 to lane 2, connecting line 3 represents the predecessor and successor information from lane 6 to lane 3, and connecting line 4 represents the predecessor and successor information from lane 7 to lane 3. It should be noted that the related attribute identification of the lane line is not represented in FIG. 2. FIG. 3 is the first road geometric information of the first road model in the embodiment of the present application, which is a simple lane line map stored in the cloud. For lanes 1-7 shown in FIG. 3, FIG. 3 is a schematic diagram of the left and right boundaries and lane line sections of the lane lines of the lanes.

[0092] In step S102, vehicle-end-detected vehicle-aware map information is obtained.

[0093] The vehicle-detected vehicle-aware map information is a local map constructed by vehicle self-perception sensors (such as lidar, camera, radar, etc.). The vehicle-aware map information includes road information such as obstacles, lane lines, road signs, pedestrians, and other vehicles around the vehicle. The vehicle perception map is usually used for local path planning, obstacle avoidance, and environmental perception.

[0094] In the embodiment, the vehicle-aware map information includes but is not limited to lane line geometric information, lane line attributes, and lane signs perceived by the vehicle sensor. Due to insufficient vehicle-end perception capability, there may be problems such as missing detection, for example, missing detection of lane signs in the lane and missing detection of lane lines. It should be noted that the vehicle-aware map information does not include vehicle predecessor and successor information at this time. Taking FIG. 4 as an example, FIG. 4 is vehicle-aware map information in the embodiment of the present application (road attributes perceived by the vehicle sensor are not represented in FIG. 4). The vehicle-aware road information of FIG. 4 includes road geometric information generated by vehicle perception, and FIG. 4 and FIG. 3 represent the same road. Compared with FIG. 3, the lane lines of lane 2 and lane 3 are missing in FIG. 4, and compared with FIG. 2, the vehicle-aware map information in FIG. 4 is also missing part of the landmark of the lane turning information in lane 1.

[0095] In step S103, the vehicle-aware map information is matched with the first road geometric information and the first road topological information of the first road model to obtain target road geometric information and target road topological information.

[0096] The automatic driving system enables real-time perception of the environment around the vehicle according to target road geometry information and target road topology information. The present application fuses the first road model in the cloud and the vehicle perception map information to obtain the target road model, so as to facilitate the automatic driving to accurately obtain global and local information. For example, the global path planning can use the cloud map, and the local obstacle avoidance relies on the vehicle map information.

[0097] In an embodiment, the matching of the vehicle perception map information with the first road geometry information and the first road topology information of the first road model to obtain target road geometry information and target road topology information comprises: matching the vehicle perception map information with the first road geometry information to obtain target road geometry information; and constructing target road topology information according to the first road topology information, the vehicle perception map information and the target road geometry information.

[0098] Due to factors such as positioning deviation of the vehicle end and cumulative perception error of the vehicle end sensor, there is a large deviation between the road model sent by the cloud and the road model constructed according to the perception of the vehicle, so it is necessary to match the geometric features of the vehicle end model first.

[0099] In an embodiment, the receiving of the vehicle perception map information generated by the vehicle end and the matching of the vehicle perception map information with the first road geometry information to obtain target road geometry information comprises: obtaining second road geometry information according to the vehicle perception map information; and performing geometric matching on the first road geometry information and the second road geometry information, and generating target road geometry information according to the matching result.

[0100] Putting the geometric structure in the cloud road structure into the vehicle end map, due to the existence of the positioning error of the vehicle and the perception error of the vehicle sensor, the vehicle perception road information of the vehicle end will have deviation when generating road geometry information. According to the real-time perception information of the vehicle end and the first road geometry information, geometric matching and calibration can be performed to fill in other geometric attributes of the lane that are not detected, such as filling in the left and right boundaries of the undetected lane.

[0101] In an embodiment, the geometric matching of the first road geometry information and the second road geometry information comprises: determining a self-perception reliable range, and determining target road geometry information according to the second road geometry information within the self-perception reliable range; and correcting the second road geometry information according to the first road geometry information to determine the target road geometry information when the self-perception reliable range is exceeded.

[0102] In the embodiment, the ego vehicle perception range can be a real vehicle sensing box. Within the real vehicle sensing box near the ego vehicle, the second road geometry information sensed by the real vehicle is more reliable, that is, the lane line information detected by the ego vehicle is consistent with the real road. In the range far from the ego vehicle beyond the real vehicle sensing box, the road geometry information in the cloud is more believed. Since the ego vehicle perception ability decreases with distance, the embodiment believes the second road geometry information within the reliable range of ego vehicle perception, thereby ensuring that the ego vehicle is safer and more reliable nearby. The first road geometry information is used to correct the second road geometry information beyond the reliable range of ego vehicle perception, thereby relying on the over-the-horizon information in the cloud to make up for the lack of vehicle perception, thereby achieving the strong combination of the map model in the ego vehicle application. In an embodiment, the ego vehicle perception range is in the distance range of 0.5 vehicle body to 1.5 vehicle body.

[0103] In an embodiment, in the case of exceeding the reliable range of ego vehicle perception, the target road geometry information is determined by correcting the second road geometry information according to the first road geometry information, including: obtaining the first road center line of the first road geometry information and the second road center line of the second road geometry information according to a preset geometry information conversion format; and weighted fitting the first road center line and the second road center line beyond the reliable range of ego vehicle perception to determine the target road geometry information.

[0104] The geometry information of the lane line is a point set in three-dimensional space. The preset geometry information conversion format is to convert the lane line point set information into a series of ordered point chain sets. Converting the point set in three-dimensional space into a point chain set includes: converting each pixel point on the lane line through the cloud vehicle coordinate system to obtain a three-dimensional coordinate point of the road, removing the height information of the three-dimensional coordinate point of the road to generate a two-dimensional data point represented by (x, y), and arranging and connecting the two-dimensional data points according to the lane line direction to generate a point chain set.

[0105] Since the lane center line geometry point chain should be located at the center of the lane line, the deviation in the geometry of the entire lane can be corrected according to the center line of the lane line. For the point chain curve 1 representing the first road center line and the point chain curve 2 representing the second road center line, a weighted transition method is used between the point chain curve 1 and the point chain curve 2, specifically including: using linear interpolation to represent the weighted transition between the point chain curve 1 and the point chain curve 2, setting a weighted average formula (1), and the formula (1) includes:

[0106] y = w(x)y1 + (1-w(x))y2 Formula (1);

[0107] where y1 and y2 are the values of the point chain curve 1 and the point chain curve 2 at this position respectively. At the starting point, w(x) takes the value of 0, at this time, the point chain curve 2 is believed 100%, as the curve moves away from the reliable range of the ego vehicle, w(x) tends to take the value of 1, and the vehicle end believes the point chain curve 1 more and more.

[0108] The specific weight function w(x) can be selected as a linear weight function as shown in formula (2), or an exponential weight function as shown in formula (3).

[0109] where L is the total length of the curve.

[0110] w(x)=e -kx formula (3);

[0111] wherein k is a positive number, used to control the decay rate of the weight, and the smoothness of the transition form.

[0112] In this embodiment, when the ego vehicle perception reliable range is exceeded, the point chain curve 1 is selected from the first road centerline set one by one, and the point chain curve 2 corresponding to the point chain curve 1 is selected from the second road centerline set one by one, and the point chain curve 1 and the point chain curve 2 are gradually fitted according to the preset weighting algorithm. Since the point chain dotted line contains road feature information, by fitting the point chain curve, the geometric features of the vehicle end map can be quickly corrected.

[0113] Since the vehicle end can only perceive part of the road topological information, for example, the vehicle end can detect the road turning information and the lane line attribute, but the vehicle end cannot directly perceive the predecessor and successor information, and the vehicle end needs to reconstruct the target road topological information on the corrected target road geometric information.

[0114] Step S103, after correcting the vehicle end road information to obtain the target road geometric information, needs to receive the first road topological information of the cloud end, and reconstruct the target road topological information according to the vehicle perception map information.

[0115] In one embodiment, the target road topological information is constructed according to the first road topological information, the vehicle perception map information and the target road geometric information, including: verifying the first road topological information, correcting the unqualified first road topological information according to the vehicle perception map information to obtain the second road topological information; comparing the vehicle perception map information and the first road topological information to determine the new road topological information; and generating the target road topological information according to the new road topological information and the second road topological information.

[0116] In this embodiment, the first topological information of the cloud end is verified first, and the vehicle perception map information collected by the vehicle end is used to correct the unreasonable first road topological information.

[0117] In an embodiment, verifying the first road topology information comprises: judging the connection rationality of the predecessor and successor information of the first road topology information according to a preset judgment rule; verifying the accuracy of the lane turning information, the predecessor and successor information, and the lane type of the first road topology information according to the vehicle sensing map information.

[0118] Since the lane turning information and the lane type are real lane attributes in the physical world and are closely bound to the landmark arrow, and are within the reliable range of self-vehicle sensing, the embodiment adopts the vehicle sensing map information to correct the lane turning information and the lane type of the first road topology information. When the reliable range of self-vehicle sensing is exceeded, the target road topology information is constructed according to the first road topology information under the condition that the first road topology information is confirmed to be reliable. It should be noted that when the vehicle sensing map information is missing, the missing part of the vehicle sensing map information is filled according to the first road topology information.

[0119] The predecessor and successor information represents the connectivity relationship and is used to represent the connectivity of the current lane and the next lane. In automatic driving, the predecessor and successor information represents the lane that has the right of way without lane changing. The network topology structure of the road can be constructed through the predecessor and successor information, thereby ensuring that at least one passable path is provided for the vehicle during automatic driving. Since the vehicle sensing map information does not include the predecessor and successor information, the target road topology model is constructed using the predecessor and successor information of the first road topology information.

[0120] When the cloud-end predecessor and successor information is used, the cloud-end predecessor and successor information needs to be verified. The preset judgment rule includes but is not limited to: evaluating the road smoothness after the connectivity of the predecessor and successor information, and / or evaluating whether the road connected by the predecessor and successor information meets the road rules, and / or whether the predecessor and successor information will pass through the boundary of the road.

[0121] In the embodiment, if the predecessor and successor information does not meet the rules, the predecessor and successor information between the front road and the rear road is regenerated according to the center line of the front road and the rear road.

[0122] In an embodiment, the new road topology information is determined by comparing the vehicle sensing map information and the first road topology information, comprising: obtaining new user driving trajectories and new user driving habits according to the vehicle sensing map information; generating the new road topology information according to the new user driving trajectories and the new user driving habits.

[0123] The embodiment collects the new driving route and the new driving behavior of the user in real time according to the vehicle feeling map information. After the new user driving track and the new user driving habit are determined, the related data is automatically returned to the full section data. After screening and checking, the new user driving track and the new user driving habit data are returned to the cloud server, and the returned data participates in the construction of the target vehicle end road model and the modification of the related road data of the cloud. The user automatically pulls the first road model containing the new data from the cloud when driving the same route next time, and starts to perform the related driving.

[0124] The embodiment converts the new user driving track and the new user driving habit into new road topology information, so that the target road model learns the driving route and the driving behavior of the user, and after all the learning is completed, the new road topology information is automatically returned to the full section data. After screening and checking, the returned data is returned to the cloud server, and subsequent cloud map construction is performed. When the user drives the same route next time, the function can be activated, and the vehicle automatically pulls the constructed map from the cloud to start driving.

[0125] In an embodiment, the new road topology information is generated according to the new user driving track and the new user driving habit, including: determining the priority of the driving lane and / or the priority of the lane-changing lane according to the new user driving track and the new user driving habit; and generating an identifier of the priority of the driving lane and / or the priority of the lane-changing lane to determine the new road topology information.

[0126] The embodiment extracts the user driving habit features from the user driving track, and can convert the user driving habit into new topology information. For example, according to the user driving track, it can be determined that the user prefers to drive in which lane and prefers to change lanes at which position in the lane. The preferred lane and the preferred lane position of the user driving are stored in the new road topology information in the form of priority identifiers, so that the subsequent automatic driving system can be more convenient according to the user habit to make navigation behavior that meets the user. In the embodiment, obtaining the user driving habit features specifically includes: driving from a starting point to a destination according to the user setting, and recording the track of the vehicle driving. The road boundary information in the learning process is recorded by using the environment perception technology, the lane preferred by the user to drive and the driving habit of the user (for example, driving on the left or on the right in the lane) are determined according to the road boundary information, the lane preferred by the user and the lane preferred by the user are marked with corresponding identifiers, and are stored in the new topology information. It should be noted that determining the lane preferred by the user to drive and the driving habit of the user can use a related AI model to learn the user track, or can determine the new user driving preference by counting the user driving track.

[0127] In an embodiment, after determining the new road topology information, the method further includes: modifying the first road topology information of the first road model according to the new road topology information.

[0128] The embodiment modifies the first road topology information, so that the first road model retains the new user driving semantics when the user pulls the first road model generated by the cloud next time.

[0129] In step S104, a target road model at the vehicle end is generated according to the target road geometry information and the target road topology information.

[0130] The target road model of the present application relies on the advantage of super-range of the cloud, first generates a set of high-quality first road model in the cloud, the first road model includes basic road geometry and topology structure, then the necessary road geometry and topology information is sent to the vehicle end, the vehicle end performs real-time geometry matching between the geometry structure and the vehicle perception map information perceived by the vehicle when consuming the prior information, thereby constructing the target road geometry feature, the target road geometry feature has more stable road structure, and the instability of the road structure caused by the large fluctuation of the vehicle perception information is weakened. At the same time, the first road topology information with prior information in the cloud is given to the road structure at the vehicle end, solving the road information loss caused by the insufficient perception ability of the vehicle end. Since the prior road structure information of the cloud is consumed in constructing the target road topology information, the memory and time resources can be saved in the construction process.

[0131] FIG. 5 is a schematic diagram of a target road model according to an embodiment of the present application. FIG. 2 is a first road model corresponding to FIG. 5, and FIG. 4 is a car sense map information corresponding to FIG. 5. In combination with FIGS. 2-5, the construction of the target road model according to the present embodiment is described. The first road geometry information in FIG. 2, i.e., FIG. 3, is obtained, and the second road geometry information in FIG. 4 is obtained. The second road geometry information and the first road geometry information are matched, the matching result is used to correct FIG. 3, and the road structure in FIG. 5 (i.e., the target road geometry information) is obtained. The target road topology information is reconstructed on the road structure in FIG. 5, which specifically includes: obtaining the first road topology information, verifying the first road topology information according to a preset rule to obtain the second road topology information, and filling the missing road turning information in the car sense map information according to the second road topology information (the left turning arrow of lane 1 in FIG. 5 is the road turning information filled by the second road information). Meanwhile, the newly added user driving trajectory and the newly added user driving habit detected by the car end are collected, and the newly added lane topology information is generated according to the newly added user driving trajectory and the newly added user driving habit, and the newly added lane topology information is added to FIG. 5 to obtain the target road model. It should be noted that the newly added user driving trajectory and the newly added user driving habit are not only applied to the construction of the target road model of the car end this time, but also used to modify or cover the user historical form trajectory and the user historical form habit stored in the cloud. Thus, when the user pulls the first road model from the cloud next time, the first road topology information of the first road model contains the newly added user driving trajectory and the newly added user driving habit, thereby ensuring the accuracy of the driving semantics of the target map of the car end next time.

[0132] The technical solutions provided by the present application can include the following beneficial effects:

[0133] In the first aspect, the present application obtains a first road model generated by the cloud and car sense map information detected by the car end. On the basis of the priori of the first road model, the car sense map information is fused to generate a target road model of the car end. Compared with the generation method of the traditional road model of the car end, the present application assigns the road information with priori information of the cloud to the road structure of the car end, saves the computing power for constructing the road model of the car end, and improves the efficiency of generating the target road model.

[0134] In the second aspect, the present application preserves the over-the-horizon advantage of the cloud in constructing the road model by consuming the first road geometry information and the first road topology information of the cloud. By matching the car sense map information, the first road geometry information, and the first road topology information, the instability of the road structure caused by the large fluctuation of the perception information of the car end can be reduced, thereby constructing a more stable target road model. Meanwhile, the car sense map information can collect the newly added user driving trajectory in real time, and the newly added user driving trajectory is fused into the target road model, so that the target road model can update the user driving features in real time.

[0135] Fig. 6 is a structural schematic diagram of a road model generation device according to an embodiment of the present application.

[0136] As shown in Fig. 6, a road model generation device comprises:

[0137] A first acquisition module 601 is configured to acquire a first road model generated in the cloud, wherein the first road model comprises at least first road geometry information and first road topology information.

[0138] A second acquisition module 602 is configured to acquire vehicle-sensed map information detected at a vehicle end.

[0139] A processing module 603 is configured to match the vehicle-sensed map information with the first road geometry information and the first road topology information of the first road model to obtain target road geometry information and target road topology information.

[0140] A determination module 604 is configured to generate a target road model at the vehicle end according to the target road geometry information and the target road topology information.

[0141] In an embodiment, the processing module further comprises a first processing module and a second processing module. The first processing module is configured to match the vehicle-sensed map information with the first road geometry information to obtain the target road geometry information. The second processing module is configured to construct the target road topology information according to the first road topology information, the vehicle-sensed map information and the target road geometry information.

[0142] In an embodiment, the first road topology information comprises at least one or more of lane turning information, predecessor and successor information, lane type and user driving trajectory.

[0143] In an embodiment, matching the vehicle-sensed map information with the first road geometry information to obtain the target road geometry information comprises: the first processing module acquires second road geometry information according to the vehicle-sensed map information; the first road geometry information and the second road geometry information are geometrically matched, and the target road geometry information is generated according to the matching result.

[0144] In an embodiment, the geometric matching of the first road geometry information and the second road geometry information comprises: the first processing module determines a self-perception reliable range, and within the self-perception reliable range, the target road geometry information is determined according to the second road geometry information; in the case of exceeding the self-perception reliable range, the second road geometry information is corrected according to the first road geometry information to determine the target road geometry information.

[0145] In an embodiment, the target road geometry information is determined by correcting the second road geometry information according to the first road geometry information when the self-perception reliable range is exceeded, including: obtaining a first road center line of the first road geometry information and a second road center line of the second road geometry information according to a preset geometry information conversion format; and fitting the first road center line and the second road center line that are outside the self-perception reliable range to determine the target road geometry information.

[0146] In an embodiment, the target road topology information is constructed according to the first road topology information, the vehicle perception map information and the target road geometry information, including: verifying the first road topology information, correcting the first road topology information that fails to pass the verification according to the vehicle perception map information to obtain second road topology information; comparing the vehicle perception map information and the first road topology information to determine new road topology information; and generating the target road topology information according to the new road topology information and the second road topology information.

[0147] In an embodiment, the first road topology information is verified, including: judging the connection rationality of the predecessor and successor information of the first road topology information according to a preset judgment rule; and / or verifying the accuracy of the lane turning information, the predecessor and successor information and the lane type of the first road topology information according to the vehicle perception map information.

[0148] In an embodiment, the new road topology information is determined by comparing the vehicle perception map information and the first road topology information, including: obtaining new user driving trajectories and new user driving habits according to the vehicle perception map information; and generating the new road topology information according to the new user driving trajectories and the new user driving habits.

[0149] FIG. 7 is a structural schematic diagram of a vehicle according to an embodiment of the present application.

[0150] Referring to FIG. 7, the vehicle 700 includes a memory 710 and a processor 720.

[0151] The processor 720 can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs) or other programmable logic devices, discrete gates or transistor logic, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor.

[0152] The memory 710 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 720 or other modules of the computer. The permanent storage device can be a read-write 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 uses a mass storage device (such as a magnetic or optical disk, flash memory) as a permanent storage device. In some other embodiments, the permanent storage device can be a removable storage device (such as a floppy disk, an optical drive). The system memory can be a read-write storage device or a volatile read-write 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 710 can include a combination of any computer readable storage media, including various types of semiconductor storage chips (such as DRAM, SRAM, SDRAM, flash memory, programmable read-only memory), magnetic disks and / or optical disks. In some embodiments, the memory 710 can include a read and / or write removable storage device, such as a compact disc (CD), a read-only digital versatile disc (such as DVD-ROM, double-layer DVD-ROM), a read-only Blu-ray disc, an ultra-density optical disc, a flash memory card (such as 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.

[0153] The executable code stored on the memory 710 can cause the processor 720 to perform part or all of the above-mentioned methods when the executable code is processed by the processor 720.

[0154] In addition, the method according to the present application can also be implemented as a computer program or 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.

[0155] 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 server, etc.), causes the processor to perform part or all of the steps of the above-mentioned methods according to the present application.

[0156] 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. The use of the terms "preferably," "preferably," "preferred," "desirably," and words of similar meaning is intended to attribute importance to a feature, method, or characteristic that addresses a problem or addresses the state of the art but is not necessarily essential to the working of the application. It is to be understood that the use of certain specific terms will not function to limit the scope of the embodiments described herein.

Claims

1. A road model generation method characterized by comprising: The method comprises the following steps: obtaining a first road model generated by a cloud, wherein the first road model comprises at least first road geometry information and first road topology information; obtaining vehicle-sensed map information detected by a vehicle; matching the vehicle-sensed map information with the first road geometry information and the first road topology information of the first road model to obtain target road geometry information and target road topology information; generating a target road model at the vehicle end according to the target road geometry information and the target road topology information.

2. The method of claim 1, wherein, The matching of the vehicle-sensed map information with the first road geometry information and the first road topology information of the first road model to obtain the target road geometry information and the target road topology information comprises: matching the vehicle-sensed map information with the first road geometry information to obtain the target road geometry information; constructing the target road topology information according to the first road topology information, the vehicle-sensed map information and the target road geometry information.

3. The method of claim 2, wherein, The first road topology information comprises at least one or more of lane turning information, predecessor and successor information, lane type and user driving trajectory.

4. The method of claim 2, wherein, The matching of the vehicle-sensed map information with the first road geometry information to obtain the target road geometry information comprises: obtaining second road geometry information according to the vehicle-sensed map information; performing geometric matching on the first road geometry information and the second road geometry information to generate the target road geometry information according to a matching result.

5. The method of claim 4, wherein, The geometric matching on the first road geometry information and the second road geometry information comprises: determining a self-perception reliable range, and determining the target road geometry information according to the second road geometry information within the self-perception reliable range; determining the target road geometry information by correcting the second road geometry information according to the first road geometry information when the self-perception reliable range is exceeded.

6. The method of claim 5, wherein, The determination of the target road geometry information by correcting the second road geometry information according to the first road geometry information when the self-perception reliable range is exceeded comprises: obtaining a first road center line of the first road geometry information and a second road center line of the second road geometry information according to a preset geometric information conversion format; performing weighted fitting on the first road center line and the second road center line that exceed the self-perception reliable range to determine the target road geometry information.

7. The method of claim 3, wherein, The construction of the target road topology information according to the first road topology information, the vehicle-sensed map information and the target road geometry information comprises: verifying the first road topology information, correcting the first road topology information that fails to pass the verification according to the vehicle-sensed map information to obtain second road topology information; comparing the vehicle-sensed map information with the first road topology information to determine new road topology information; generating the target road topology information according to the new road topology information and the second road topology information.

8. The method of claim 7, wherein, The verification of the first road topology information comprises: judging the connection rationality of the predecessor and successor information of the first road topology information according to a preset judgment rule; and / or According to the vehicle map information, the accuracy of lane turning information, front and rear information and lane type of the first road topology information is verified.

9. The method of claim 7, wherein, The comparison of the vehicle map information and the first road topology information determines the new road topology information, including: According to the vehicle map information, the new user driving track and the new user driving habit are obtained; According to the new user driving track and the new user driving habit, the new road topology information is generated.

10. A road model generation device characterized by comprising: Including: The first acquisition module is used for acquiring the first road model generated by the cloud, and the first road model at least includes first road geometry information and first road topology information; The second acquisition module is used for acquiring the vehicle map information detected by the vehicle side; The processing module is used for matching the vehicle map information with the first road geometry information and the first road topology information of the first road model to obtain target road geometry information and target road topology information; The determination module generates the target road model of the vehicle side according to the target road geometry information and the target road topology information.

11. The apparatus of claim 10, wherein, The processing module includes: The first processing module is used for matching the vehicle map information with the first road geometry information to obtain target road geometry information; The second processing module is used for constructing target road topology information according to the first road topology information, the vehicle map information and the target road geometry information.

12. The apparatus of claim 11, wherein: The first processing module acquires second road geometry information according to the vehicle map information; The first road geometry information and the second road geometry information are geometrically matched, and target road geometry information is generated according to the matching result.

13. The apparatus of claim 12, wherein: The first processing module determines a self-vehicle perception reliable range, and within the self-vehicle perception reliable range, target road geometry information is determined according to the second road geometry information; In the case of exceeding the self-vehicle perception reliable range, the second road geometry information is corrected according to the first road geometry information to determine the target road geometry information.

14. An automobile characterized by comprising: Including: A processor; And A memory having executable code stored thereon, which, when executed by the processor, causes the processor to perform the method of any one of claims 1 to 9.

15. A computer-readable storage medium, characterized in that, The memory having executable code stored thereon, which, when executed by the processor of the electronic device, causes the processor to perform the method of any one of claims 1-9.

Images