Vehicle control method and device

Abstract

The invention relates to a vehicle control method, and the method comprises the steps: detecting that a lane planned by a vehicle is blocked by other obstacles in a parking lot scene; extracting a local high-precision map from a predetermined parking lot high-precision map according to the pose of the vehicle, and converting the local high-precision map into an occupied grid map; fusing the occupied grid map with a map perceived by the vehicle in real time so as to obtain an updated grid map; in the updated grid map, constructing a virtual lane section to replace the blocked lane section; and controlling the vehicle based on the virtual lane segment. The application also relates to a vehicle control device, a computer readable storage medium, a computer program product and an automatic parking system.

Description

Technical Field

[0001] The present application relates to the field of vehicle control, and more specifically, to a vehicle control method and device, a computer-readable storage medium, a computer program product, and an automatic parking system. Background Art

[0002] Existing automatic parking systems generally plan driving lanes only based on the marked real roads. When the lane is blocked by a vehicle or other obstacles, the automatic parking system will require the driver to take over the vehicle or be directly trapped there. However, in actual scenarios (such as a parking lot), there may be free space to continue moving forward.

[0003] For example, referring to Figure 4 , it shows a top view of a parking lot scenario, where the vehicle 410 is driving along the lane planned by the autonomous driving system, but the lane is blocked by other lanes. At this time, if the free space in the parking lot (such as the unoccupied parking space shown in Figure 4 ) can be fully utilized, for example, re-planning the path as shown by 420, the obstacles in front of the vehicle can be avoided, saving time and effort for the driver and improving satisfaction. Summary of the Invention

[0004] According to one aspect of the present application, there is provided a vehicle control method, the method including: in a parking lot scenario, detecting that the lane planned by the vehicle is blocked by other obstacles; extracting a local high-precision map from a pre-determined high-precision map of the parking lot according to the pose of the vehicle, and converting the local high-precision map into an occupancy grid map; fusing the occupancy grid map with the map sensed by the vehicle in real time to obtain an updated grid map; constructing a virtual lane segment in the updated grid map to replace the blocked lane segment; and controlling the vehicle based on the virtual lane segment.

[0005] As a supplement or replacement to the above solution, the above method may further include: sending the virtual lane segment to a human-machine interface to inform the driver.

[0006] As a supplement or replacement to the above solution, in the above method, the planned lane is planned by the vehicle's autonomous driving system or assisted driving system based on the high-precision map of the parking lot.

[0007] As a supplement or replacement to the above solution, in the above method, according to the pose of the vehicle, a local high-precision map is extracted from a pre-determined high-precision parking lot map, and converting the local high-precision map into an occupancy grid map includes: extracting a local area map with the vehicle position as the origin from the high-precision parking lot map; meshing the local area map; and determining whether the elements in the local area map are occupied according to the semantic elements in the local area map, so as to generate the occupancy grid map.

[0008] As a supplement or replacement to the above solution, in the above method, fusing the occupancy grid map with the map sensed by the vehicle in real time to obtain an updated grid map includes: using the obstacle information sensed by the vehicle in real time to update the occupancy information of the occupancy grid map, so as to generate the updated grid map.

[0009] As a supplement or replacement to the above solution, in the above method, in the updated grid map, constructing a virtual lane segment to replace the blocked lane segment includes: searching for the shortest path connecting to the next target point that is not blocked by obstacles in the planned lane in the updated grid map; and using the shortest path as the center line to generate the virtual lane segment to replace the blocked lane segment.

[0010] According to another aspect of the present application, a vehicle control device is provided, the device includes: a detection device, configured to detect that the lane planned by the vehicle is blocked by other obstacles in a parking lot scenario; a conversion device, configured to extract a local high-precision map from a pre-determined high-precision parking lot map according to the pose of the vehicle, and convert the local high-precision map into an occupancy grid map; a fusion device, configured to fuse the occupancy grid map with the map sensed by the vehicle in real time to obtain an updated grid map; a construction device, configured to construct a virtual lane segment to replace the blocked lane segment in the updated grid map; and a control device, configured to control the vehicle based on the virtual lane segment.

[0011] As a supplement or replacement to the above solution, the above device may further include: a sending device, configured to send the virtual lane segment to a human-machine interface to inform the driver.

[0012] As a supplement or replacement to the above solution, in the above device, the conversion device is configured to: extract a local area map with the vehicle position as the origin from the high-precision parking lot map; mesh the local area map; and determine whether the elements in the local area map are occupied according to the semantic elements in the local area map, so as to generate the occupancy grid map.

[0013] As a supplement or replacement to the above solution, in the above device, the fusion device is configured to: update the occupancy information of the occupancy grid map by using the obstacle information sensed by the vehicle in real time, so as to generate the updated grid map.

[0014] As a supplement or replacement to the above solution, in the above device, the construction device is configured to: search for the shortest path connecting to the next target point that is not blocked by obstacles in the planned lane in the updated grid map; and generate the virtual lane segment with the shortest path as the center line to replace the blocked lane segment.

[0015] According to another aspect of the present application, there is provided a computer-readable storage medium, the medium includes instructions, and the instructions execute the method as described above when running.

[0016] According to another aspect of the present application, there is provided a computer program product, including a computer program, and when the computer program is executed by a processor, the method as described above is implemented.

[0017] According to another aspect of the present application, there is provided an automatic parking system, and the automatic parking system includes the vehicle control device as described above.

[0018] The vehicle control solution of the embodiment of the present application, when the lane planned by the vehicle is blocked by other obstacles, generates a virtual lane segment by using the high-precision map (semantic vectorized map) of the parking lot, avoiding manual takeover when the marked or planned lane is blocked, saving time and effort for the driver and improving satisfaction. In one embodiment, the vehicle control solution of the embodiment of the present application further ensures the driving safety of the generated virtual lane segment by updating the occupancy information of the occupancy grid map by using the obstacle information sensed by the vehicle in real time. In one embodiment, the vehicle control solution of the embodiment of the present application sends the generated virtual lane segment to the human-machine interface HMI (such as the central control screen) to notify the driver in a visual manner in a timely manner. Description of the Drawings

[0019] From the following detailed description in conjunction with the drawings, the above and other objects and advantages of the present application will become more completely clear, wherein the same or similar elements are denoted by the same reference numerals.

[0020] Figure 1 The flowchart of the vehicle control method according to an embodiment of the present application is shown;

[0021] Figure 2 The structural diagram of the vehicle control device according to an embodiment of the present application is shown;

[0022] Figure 3The figure shows a schematic flow diagram of a vehicle control method according to an embodiment of the present application; and

[0023] Figure 4 The figure shows a schematic diagram of generating a virtual lane segment in a parking lot scenario according to an embodiment of the present application. Detailed implementation manners

[0024] Hereinafter, vehicle control solutions according to various exemplary embodiments of the present application will be described in detail with reference to the accompanying drawings.

[0025] Figure 1 The figure shows a schematic flow diagram of a vehicle control method 1000 according to an embodiment of the present application. As Figure 1 shown, the vehicle control method 1000 includes:

[0026] In step S110, in a parking lot scenario, it is detected that the lane planned by the vehicle itself is blocked by other obstacles;

[0027] In step S120, according to the pose of the vehicle itself, a local high-precision map is extracted from a pre-determined high-precision parking lot map, and the local high-precision map is converted into an occupancy grid map;

[0028] In step S130, the occupancy grid map is fused with the map sensed by the vehicle in real time to obtain an updated grid map;

[0029] In step S140, in the updated grid map, a virtual lane segment is constructed to replace the blocked lane segment; and

[0030] In step S150, the vehicle itself is controlled based on the virtual lane segment.

[0031] In one or more embodiments, the vehicle control method 1000 of the present application is executed by an automatic parking system in a parking lot scenario. Among them, in step S110, in a parking lot scenario, it is detected that the lane planned by the vehicle itself is blocked by other obstacles. In one or more embodiments, the planned lane (or trajectory) is planned by the intelligent driving system of the vehicle itself based on a high-precision parking lot map (provided by a map provider or created by the intelligent driving system itself), where the high-precision map contains semantic information.

[0032] Moreover, since the high-precision map can provide static semantic elements, some short-term static elements or dynamic elements (such as vehicles) generally do not belong to the high-precision map. Therefore, the lane planned only based on the high-precision parking lot map may be blocked by other obstacles (such as other vehicles) due to traffic accidents, etc.

[0033] In step S120, according to the pose of the vehicle, a local high-precision map is extracted from a pre-determined high-precision parking lot map, and the local high-precision map is converted into an occupancy grid map. In one embodiment, step S120 includes: extracting a local area map with the vehicle position as the origin from the high-precision parking lot map; meshing the local area map; and determining whether the elements are occupied according to the semantic elements in the local area map, so as to generate the occupancy grid map.

[0034] In one or more embodiments, the local area map with the vehicle position as the origin is a map of the local area around the vehicle (such as 15 meters in front of the vehicle, 5 meters on the left and right, and 5 meters behind). In this way, according to the pose of the vehicle (including position and orientation), a local area map with the vehicle position as the origin can be extracted from the high-precision parking lot map.

[0035] In one or more embodiments, meshing the local area map uses grids of equal size, such as 5cm x 5cm.

[0036] In addition, the semantic elements in the high-precision map have various types. For example, some semantic elements contain ground markings: such as lane lines, guiding arrows, zebra crossings, parking space lines, etc. The ground represented by these ground markings has no height, so it is non-occupied from a semantic perspective. In addition, there are some other semantic elements, such as pillars in the parking lot, ground wheel stops, parking space locks, etc., which represent semantic elements with height. Therefore, in the context of this application, the positions of these elements are occupied.

[0037] In step S130, the occupancy grid map is fused with the map sensed by the vehicle in real time to obtain an updated grid map. In one embodiment, step S130 includes: using the obstacle information sensed by the vehicle in real time to update the occupancy information of the occupancy grid map, so as to generate the updated grid map. As mentioned above, some short-term static elements generally do not belong to the map. In a parking lot, for example, vehicles parked in parking spaces or other stationary obstacles. These obstacles can generally be represented in the vehicle coordinate system in the sensing system and can be conveniently converted into the grid map to update the occupancy information of the grid map.

[0038] In step S140, in the updated grid map, a virtual lane segment is constructed to replace the blocked lane segment. In one embodiment, step S140 may include: in the updated grid map, various search algorithms (including but not limited to the A* algorithm) can be used to search for the shortest path connecting to the next target point that is not blocked by obstacles in the planned lane; and using the shortest path as the center line, generating the virtual lane segment to replace the blocked lane segment.

[0039] In one embodiment, although Figure 1 not shown in the figure, the above method 1000 may further include: sending the virtual lane segment to a human-machine interface HMI to inform the driver.

[0040] The foregoing vehicle control method 1000 generates a virtual lane segment by using a high-precision parking lot map (semantic vectorized map) when the lane planned by the vehicle itself is blocked by other obstacles, avoiding manual takeover when the marked or planned lane is blocked, saving time and effort for the driver, and improving satisfaction.

[0041] Figure 3 The flowchart of a vehicle control method 3000 according to an embodiment of the present application is shown. In Figure 3 the embodiment, it is assumed that a high-precision semantic vectorized map has been prefabricated. In step S310, the vehicle is driven by an automatic driving program in a parking lot, where the vehicle travels normally according to the lanes drawn on the high-precision parking lot map. Then, in step S320, the lane blockage detection function of the vehicle is used to detect whether the planned path is blocked. If not, in step S325, the vehicle will travel normally. If so, that is, the planned lane is blocked, step S330 is executed. In step S330, according to the pose of the vehicle, a local high-precision semantic vectorized map is extracted and then converted into an occupancy grid map (OGM). Then, in step S340, the OGM is fused with the real-time local perception input. Subsequently, in step S350, a short path is searched to connect to the next unblocked target point. Various search algorithms can be used, including the A* algorithm or other variants. Taking this searched path as the center line, a virtual lane segment can be generated to replace the blocked lane segment. Then, in step S360, it is judged whether the virtual lane segment is found. If not, step S365 is executed, and the automatic driving system will pause the vehicle and then request the driver to take over. If so, that is, the virtual lane segment is found, steps S370 and S375 are executed. In step S370, the virtual lane segment is sent to the HMI module and displayed to the driver on the in-vehicle display screen. In step S375, the virtual lane segment is also sent to the planner module, thereby planning a local trajectory as a reference trajectory for controlling the vehicle.

[0042] In addition, those skilled in the art can easily understand that the vehicle control method 1000 provided by one or more of the above embodiments of the present application can be implemented by a computer program. For example, the computer program is included in a computer program product, and when the computer program is executed by a processor, it implements the vehicle control method 1000 of one or more embodiments of the present application. For another example, when a computer-readable storage medium (such as a USB flash drive) storing the computer program is connected to a computer, running the computer program can execute the vehicle control method 1000 of one or more embodiments of the present application.

[0043] Reference Figure 2 , which shows a schematic structural diagram of a vehicle control device 2000 according to an embodiment of the present application. As Figure 2 shown, the vehicle control device 2000 includes a detection device 210, a conversion device 220, a fusion device 230, a construction device 240, and a control device 250. Among them, the detection device 210 is used to detect that the lane planned by the vehicle itself is blocked by other obstacles in a parking lot scenario; the conversion device 220 is used to extract a local high-precision map from a pre-determined parking lot high-precision map according to the pose of the vehicle itself, and convert the local high-precision map into an occupancy grid map; the fusion device 230 is used to fuse the occupancy grid map with the map sensed by the vehicle in real time to obtain an updated grid map; the construction device 240 is used to construct a virtual lane segment in the updated grid map to replace the blocked lane segment; and the control device 250 is used to control the vehicle based on the virtual lane segment.

[0044] In one or more embodiments, the planned lane (or trajectory) is planned by the vehicle's intelligent driving system based on a parking lot high-precision map (provided by a map provider or created by the intelligent driving system itself), where the high-precision map contains semantic information.

[0045] Moreover, since the high-precision map can provide static semantic elements, some short-term static elements or dynamic elements (such as vehicles) generally do not belong to the high-precision map. Therefore, the lane planned only based on the parking lot high-precision map may be blocked by other obstacles (such as other vehicles) due to traffic accidents or the like.

[0046] The conversion device 220 is used to extract a local high-precision map from a pre-determined parking lot high-precision map according to the pose of the vehicle itself, and convert the local high-precision map into an occupancy grid map. In one embodiment, the conversion device 220 is configured to: extract a local area map with the vehicle position as the origin from the parking lot high-precision map; grid the local area map; and determine whether the element is occupied according to the semantic elements in the local area map, so as to generate the occupancy grid map.

[0047] In one or more embodiments, the local area map with the vehicle position as the origin is a map of the local area around the vehicle (such as 15 meters in front of the vehicle, 5 meters on the left and right, and 5 meters behind). In this way, according to the pose of the vehicle (including position and orientation), the conversion device 220 can extract a local area map with the vehicle position as the origin from the high-precision parking lot map. In one or more embodiments, the conversion device 220 meshes the local area map using equally sized grids, such as 5cm x 5cm. Additionally, there are various types of semantic elements in the high-precision map. For example, some semantic elements include ground markings: such as lane lines, guiding arrows, zebra crossings, parking space lines, etc. The ground representations represented by these ground markings have no height and are therefore non-occupied from a semantic perspective. There are also some other semantic elements, such as the pillars in the parking lot, ground wheel stops, parking space locks, etc., which represent semantic elements with height. Therefore, in the context of this application, the positions of these elements are occupied.

[0048] The fusion device 230 is used to fuse the occupied grid map with the map of the vehicle's real-time perception to obtain an updated grid map. In one embodiment, the fusion device 230 is configured to: use the obstacle information of the vehicle's real-time perception to update the occupancy information of the occupied grid map, thereby generating the updated grid map. As mentioned before, there are some short-term static elements that generally do not belong to the map. In a parking lot, for example, vehicles parked in parking spaces or other stationary obstacles. These obstacles can generally be represented in the vehicle's coordinate system in the perception system and can be conveniently converted into the grid map. The fusion device 230 updates the occupancy information of the grid map based on this.

[0049] The construction device 240 is used to construct virtual lane segments in the updated grid map to replace the blocked lane segments. In one embodiment, the construction device 240 is configured to: in the updated grid map, use a search algorithm (including but not limited to the A* algorithm) to search for the shortest path connecting to the next target point in the planned lane that is not blocked by obstacles; and use the shortest path as the center line to generate the virtual lane segment to replace the blocked lane segment.

[0050] In one embodiment, although Figure 2 not shown in the figure, the above device 2000 may further include: a sending device for sending the virtual lane segment to the human-machine interface HMI to inform the driver.

[0051] Through the aforementioned vehicle control device 2000, when the lane planned by the vehicle is blocked by other obstacles, a virtual lane segment can be generated using the high-precision parking lot map (semantic vectorized map), for example, as Figure 4As shown in 420, it bypasses the blocked section and avoids unnecessary manual takeover, saving time and effort for the driver and improving satisfaction.

[0052] In one or more embodiments, the above vehicle control device 2000 can be integrated into various types of automatic parking systems, including but not limited to, autonomous valet parking (AVP), memory parking (HPA), etc.

[0053] In summary, the vehicle control solution of the embodiments of the present application generates a virtual lane segment by using the high-precision map (semantic vectorized map) of the parking lot when the lane planned by the vehicle itself is blocked by other obstacles, avoiding manual takeover when the marked or planned lane is blocked, saving time and effort for the driver and improving satisfaction. In one embodiment, the vehicle control solution of the embodiments of the present application can further ensure the driving safety of the generated virtual lane segment by updating the occupancy information of the occupancy grid map by using the obstacle information sensed by the vehicle in real time. In one embodiment, the vehicle control solution of the embodiments of the present application sends the generated virtual lane segment to the human-machine interface HMI (such as the center console screen) to notify the driver in a visual manner in a timely manner

[0054] The above examples mainly illustrate the vehicle control solution of the embodiments of the present application. Although only some of the embodiments of the present application are described, those of ordinary skill in the art should understand that the present application can be implemented in many other forms without departing from its gist and scope. Therefore, the examples and embodiments shown are regarded as illustrative rather than restrictive, and the present application may cover various modifications and substitutions without departing from the spirit and scope of the present application as defined by the various claims.

Claims

1. A vehicle control method, characterized in that, The method includes: In a parking lot scenario, detecting that the lane planned by the vehicle itself is blocked by other obstacles; According to the pose of the vehicle itself, extracting a local high-precision map from a pre-determined parking lot high-precision map, and converting the local high-precision map into an occupancy grid map; Fusing the occupancy grid map with the map sensed by the vehicle itself in real time to obtain an updated grid map; In the updated grid map, constructing a virtual lane segment to replace the blocked lane segment; and Controlling the vehicle itself based on the virtual lane segment.

2. The method according to claim 1, further comprising: Sending the virtual lane segment to a human-machine interface to inform the driver.

3. The method according to claim 1, wherein, The planned lane is planned by the vehicle's autonomous driving system or assisted driving system based on the parking lot high-precision map.

4. The method according to claim 1, wherein, According to the pose of the vehicle itself, extracting a local high-precision map from a pre-determined parking lot high-precision map, and converting the local high-precision map into an occupancy grid map includes: Extracting a local area map with the vehicle position as the origin from the parking lot high-precision map; Meshing the local area map; and According to the semantic elements in the local area map, determining whether the elements are occupied, thereby generating the occupancy grid map.

5. The method according to claim 1, wherein, Fusing the occupancy grid map with the map sensed by the vehicle itself in real time to obtain an updated grid map includes: Using the obstacle information sensed by the vehicle itself in real time to update the occupancy information of the occupancy grid map, thereby generating the updated grid map.

6. The method according to claim 1, wherein In the updated grid map, constructing a virtual lane segment to replace the blocked lane segment includes: In the updated grid map, searching for the shortest path connecting to the next target point in the planned lane that is not blocked by obstacles; and Using the shortest path as the center line to generate the virtual lane segment to replace the blocked lane segment.

7. A vehicle control device, characterized in that, The device includes: A detection device for detecting, in a parking lot scenario, that the lane planned by the vehicle itself is blocked by other obstacles; A conversion device for extracting a local high-precision map from a pre-determined parking lot high-precision map according to the pose of the vehicle itself, and converting the local high-precision map into an occupancy grid map; A fusion device for fusing the occupancy grid map with the map sensed by the vehicle itself in real time to obtain an updated grid map; A construction device for constructing a virtual lane segment to replace the blocked lane segment in the updated grid map; and A control device for controlling the vehicle itself based on the virtual lane segment.

8. The device according to claim 7, further comprising: A sending device for sending the virtual lane segment to a human-machine interface to inform the driver.

9. The device according to claim 7, wherein, The conversion device is configured to: Extract a local area map with the vehicle position as the origin from the parking lot high-precision map; Mesh the local area map; and According to the semantic elements in the local area map, determine whether the elements are occupied, thereby generating the occupancy grid map.

10. The apparatus according to claim 7, wherein, The fusion device is configured to: Use the obstacle information sensed by the vehicle itself in real time to update the occupancy information of the occupancy grid map, thereby generating the updated grid map.

11. The device according to claim 7, wherein, The construction device is configured to: In the updated grid map, search for the shortest path connecting to the next target point that is not blocked by obstacles in the planned lane; and Taking the shortest path as the center line, generate the virtual lane segment to replace the blocked lane segment.

12. A computer-readable storage medium, characterized in that, The medium includes instructions that, when running, execute the method according to any one of claims 1 to 6.

13. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the method according to any one of claims 1 to 6.

14. An automatic parking system, characterized in that, The system includes the device according to any one of claims 7 to 11.

Images