Reference line determination method, device, equipment, medium and product

Abstract

The present application discloses a reference line determination method, device, equipment, medium and product. The method includes: obtaining target vehicle data corresponding to the target vehicle, the target vehicle data including target environment data and target driving data; determining the scene in which the target vehicle is located based on the target vehicle data; when the scene is a target scene, determining multiple target control points based on the control point determination method corresponding to the target scene; creating a first reference line based on the multiple target control points; adjusting the second reference line based on the first reference line to obtain a target reference line, where the second reference line is obtained by smoothing the lane centerline. In this way, the target reference line is more suitable for the scene in which the target vehicle is currently located, and the driving trajectory generated based on the target reference line is more consistent with the driving trajectory of a human-driven vehicle, thereby improving the intelligence of the autonomous driving vehicle.

Description

Technical Field

[0001] The present application relates to the field of autonomous driving technology, and in particular to a reference line determination method, device, equipment, medium and product. Background Art

[0002] With the rapid development of science and technology and the continuous improvement of living standards, the application of autonomous driving technology is becoming more and more common.

[0003] In existing technologies, reference lines are typically generated directly from lane centerlines in high-precision maps, and the autonomous driving trajectory is then determined based on these reference lines. However, in some real-world scenarios, these reference lines are not ideal. For example, at an intersection with a turn zone, if the traffic light is green, a human-driven vehicle turning left typically does not follow the centerline of the turn zone, but instead chooses a shorter, faster path. However, an autonomous vehicle determining its trajectory based on a reference line generated from the lane centerline will follow the centerline of the turn zone.

[0004] Therefore, in some practical scenarios, the reference line generated based on the lane centerline is not reasonable, which will lead to poor intelligence of the autonomous driving vehicle. Summary of the Invention

[0005] The embodiments of the present application provide a reference line determination method, apparatus, equipment, medium, and product, which can make the target reference line more suitable for the scene in which the target vehicle is currently located. The driving trajectory generated based on the target reference line is more consistent with the driving trajectory of a human-driven vehicle, thereby improving the intelligence of the autonomous driving vehicle.

[0006] In a first aspect, an embodiment of the present application provides a reference line determination method, the method comprising:

[0007] Obtain target vehicle data corresponding to the target vehicle, the target vehicle data including target environment data and target driving data;

[0008] Determine the scene where the target vehicle is located based on the target vehicle data;

[0009] In the case where the scene is a target scene, a plurality of target control points are determined according to a control point determination method corresponding to the target scene;

[0010] creating a first reference line based on a plurality of target control points;

[0011] The second reference line is adjusted based on the first reference line to obtain a target reference line, where the second reference line is obtained by smoothing the lane center line.

[0012] In a second aspect, an embodiment of the present application provides a reference line determination device, the device comprising:

[0013] An acquisition module is used to acquire target vehicle data corresponding to a target vehicle, the target vehicle data including target environment data and target driving data;

[0014] A first determining module is used to determine the scene where the target vehicle is located based on the target vehicle data;

[0015] A second determination module is configured to determine a plurality of target control points according to a control point determination method corresponding to the target scene when the scene is a target scene;

[0016] A creation module, configured to create a first reference line based on a plurality of target control points;

[0017] The adjustment module is used to adjust the second reference line based on the first reference line to obtain a target reference line, where the second reference line is obtained by smoothing the lane center line.

[0018] In a third aspect, an embodiment of the present application provides an electronic device, comprising: a processor and a memory storing computer program instructions;

[0019] When the processor executes the computer program instructions, the reference line determination method as shown in any one of the embodiments of the first aspect is implemented.

[0020] In a fourth aspect, an embodiment of the present application provides a computer storage medium having computer program instructions stored thereon. When the computer program instructions are executed by a processor, the reference line determination method shown in any one of the embodiments of the first aspect is implemented.

[0021] In a fifth aspect, an embodiment of the present application provides a computer program product. When the instructions in the computer program product are executed by a processor of an electronic device, the electronic device executes the reference line determination method shown in any one of the embodiments of the first aspect.

[0022] The reference line determination method, device, equipment, medium and product of the embodiments of the present application can obtain target vehicle data corresponding to the target vehicle, and the target vehicle data includes target environment data and target driving data, and then determine the scene in which the target vehicle is located based on the target vehicle data. In the case where the scene is a target scene, multiple target control points are determined according to the control point determination method corresponding to the target scene, and then a first reference line is created based on the multiple target control points. Then, the second reference line is adjusted based on the first reference line to obtain a target reference line, and the second reference line is obtained by smoothing the center line of the lane. In other words, the first reference line can be determined based on the specific scene in which the target vehicle is located, and the second reference line obtained by directly smoothing the center line of the lane can be adjusted based on the first reference line to obtain the target reference line. In this way, the target reference line can be made more suitable for the scene in which the target vehicle is currently located, and the driving trajectory generated based on the target reference line is more consistent with the driving trajectory of a human-driven vehicle, thereby improving the intelligence of the autonomous driving vehicle. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0024] Figure 1 This is a flowchart of a reference line determination method provided by one embodiment of the present application;

[0025] Figure 2 This is a reference line diagram provided by an embodiment of the present application;

[0026] Figure 3 This is another reference line schematic diagram provided by an embodiment of the present application;

[0027] Figure 4 This is another reference line schematic diagram provided by an embodiment of the present application;

[0028] Figure 5 This is another reference line schematic diagram provided by an embodiment of the present application;

[0029] Figure 6 This is a schematic structural diagram of a reference line determination device provided by an embodiment of the present application;

[0030] Figure 7 This is a schematic structural diagram of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0031] The features and exemplary embodiments of various aspects of the present application will be described in detail below. In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain the present application, rather than to limit the present application. For those skilled in the art, the present application can be implemented without the need for some of these specific details. The following description of the embodiments is merely to provide a better understanding of the present application by illustrating the examples of the present application.

[0032] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, the elements defined by the phrase "comprising..." do not exclude the presence of other identical elements in the process, method, article, or device comprising the elements.

[0033] As mentioned in the background technology, existing autonomous driving technologies usually rely heavily on high-precision maps. High-precision maps provide lane centerlines that are used to generate reference lines. However, for areas without lanes, such as intersections, virtual lanes and virtual lane centerlines also exist in high-precision maps. These lane centerlines are fixed when the autonomous driving program is started. However, some reference lines generated based on lane centerlines are not reasonable. Different optimal reference lines should be used based on different scenarios. For example, in a left-turn scenario at an intersection with a turn area, when the traffic light is red, the reference line should be closely aligned with the lane centerline of the turn area and smoothly extend to the target lane. However, when the light is green, manual driving will not strictly enforce it and will choose a shorter and more convenient path. In addition, the lane centerlines depicted in some high-precision maps are intelligently collected, which is not reasonable. Faced with massive scenes, the workload of manual modification is large and the process is complicated, which is not in line with the direction of social development.

[0034] Furthermore, existing autonomous driving algorithms rely heavily on lane centerlines, using them to smooth out reference lines and then optimizing driving trajectories based on these lines. This process lacks overall flexibility, reducing vehicle intelligence and passenger comfort.

[0035] The embodiments of the present application provide a reference line determination method, apparatus, equipment, medium and product, which can obtain target vehicle data corresponding to the target vehicle, the target vehicle data including target environment data and target driving data, and then determine the scene in which the target vehicle is located based on the target vehicle data. In the case where the scene is a target scene, multiple target control points are determined according to the control point determination method corresponding to the target scene, and then a first reference line is created based on the multiple target control points. Then, a second reference line is adjusted based on the first reference line to obtain a target reference line, and the second reference line is obtained by smoothing the center line of the lane. In other words, the first reference line can be determined based on the specific scene in which the target vehicle is located, and the second reference line obtained by directly smoothing the center line of the lane can be adjusted based on the first reference line to obtain the target reference line. In this way, the target reference line can be made more suitable for the scene in which the target vehicle is currently located, and the driving trajectory generated based on the target reference line is more consistent with the driving trajectory of a human-driven vehicle, thereby improving the intelligence of the autonomous driving vehicle.

[0036] Because the vehicle's intelligence is already enhanced when the reference line is determined, there's no need to adjust the vehicle's intelligence when subsequently determining the driving trajectory based on the reference line, effectively separating the two goals of improving vehicle intelligence from improving ride comfort. Furthermore, there's no need to manually modify the lane centerline depicted on the HD map.

[0037] Figure 1 FIG. 1 is a flow chart of a reference line determination method provided by an embodiment of the present application. It should be noted that the execution subject of the reference line determination method may be a reference line determination device, such as Figure 1 As shown, the reference line determination method may include the following steps:

[0038] S110, obtaining target vehicle data corresponding to the target vehicle;

[0039] S120, determining the scene where the target vehicle is located based on the target vehicle data;

[0040] S130, when the scene is a target scene, determining a plurality of target control points according to a control point determination method corresponding to the target scene;

[0041] S140, creating a first reference line based on the multiple target control points;

[0042] S150: Adjust the second reference line based on the first reference line to obtain a target reference line.

[0043] In this way, the target vehicle data corresponding to the target vehicle can be obtained, and the target vehicle data includes target environment data and target driving data. Then, the scene in which the target vehicle is located can be determined based on the target vehicle data. In the case where the scene is a target scene, multiple target control points are determined according to the control point determination method corresponding to the target scene. Then, a first reference line is created based on the multiple target control points. Then, based on the first reference line, a second reference line is adjusted to obtain a target reference line. The second reference line is obtained by smoothing the center line of the lane. In other words, the first reference line can be determined based on the specific scene in which the target vehicle is located, and the second reference line obtained by directly smoothing the center line of the lane can be adjusted based on the first reference line to obtain the target reference line. In this way, the target reference line can be made more suitable for the scene in which the target vehicle is currently located, and the driving trajectory generated based on the target reference line is more consistent with the driving trajectory of a human-driven vehicle, thereby improving the intelligence of the autonomous driving vehicle.

[0044] Regarding S110 , the target vehicle may be an autonomous driving vehicle. The target vehicle data may include target environment data and target driving data.

[0045] In some embodiments, to more accurately determine the scene in which the target vehicle is located, the target environment data may include traffic signal data, road data, obstacle data, and a first speed and a first position of a first vehicle whose distance from the target vehicle does not exceed a distance threshold; the road data may include a lane position, lane line data, a point sequence of a lane centerline, and lane turning attributes;

[0046] The target driving data may include a target speed, a target acceleration, and a target position of the target vehicle.

[0047] Specifically, the first vehicle may include all vehicles whose distance from the target vehicle does not exceed a distance threshold, the first speed may be the current speed of the first vehicle, and the first position may be the current position of the first vehicle. The target speed may be the current speed of the target vehicle, the target acceleration may be the current acceleration of the target vehicle, and the target position may be the current position of the target vehicle.

[0048] Here, the target environmental data may be environmental data collected during the target vehicle's travel. The target environmental data can be acquired through sensors on the target vehicle. For example, a laser radar (LiDAR) can be used to acquire 3D point cloud information of the road, or a camera can be used to acquire an RGB image of the road. The 3D point cloud information may include obstacle data, the first speed and first position of the first vehicle, and the RGB image may include traffic signal data. Furthermore, road data may be obtained from high-precision maps or RGB images. The target vehicle's control system may acquire the target travel data.

[0049] Among them, traffic signal data can be traffic signal light data such as red light, green light, and yellow light. Lane line data can be lane line data of the lane where the target vehicle is currently located. The lane line data can include but is not limited to lane boundary attributes, lane boundary lines, and the connection relationship between the lane where the target vehicle is currently located and the front and rear lanes. Obstacle data can include but is not limited to whether the obstacle is stationary or moving, the speed of the obstacle, the acceleration of the obstacle, the position of the obstacle, and the shape of the obstacle. Lane turning attributes can include left-turn lanes, right-turn lanes, etc., and of course can also include other lane attributes, which are not limited here. The point sequence of the lane centerline can be a vector composed of multiple points, and the points can be represented by coordinates in the three directions of x, y, and z. The first speed can be the speed of the first vehicle located near the target vehicle.

[0050] In this way, through the above-mentioned multiple target environment data and target driving data, the scene in which the target vehicle is located can be determined more comprehensively and accurately.

[0051] Regarding S120 , based on the target vehicle data, that is, the target environment data and the target driving data, the scene in which the target vehicle is currently located can be determined.

[0052] In some implementations, the scene in which the target vehicle is located may be determined based on the target vehicle data using preset rules.

[0053] Here, the target vehicle data may include the vertical distance between the end point of the current lane segment and the previous lane segment, the curvature of the current lane segment, the steering properties of the current lane segment, static obstacle information, road vehicle information, intersection condition information of the current lane segment (whether it is at an intersection), etc.

[0054] The preset rule may be a preset correspondence between vehicle information and scenes.

[0055] For example, by traversing the lane segment sequence, the vertical distance between the endpoint of the current lane segment and the previous lane segment, as well as the first distance between the projection of the endpoint on the previous lane segment and the starting point of the previous lane segment, can be calculated. If the ratio of this vertical distance to the first distance is greater than a threshold, and the current lane segment's turning attribute is a left turn, the current lane segment has intersection information, the stop line is not at the starting point of the current lane segment, the traffic light is green, and the curvature of the current lane segment exceeds a curvature threshold, then the target vehicle can be determined to be in a green light scenario with a left turn waiting area.

[0056] In some implementations, the above S120 may include:

[0057] The target vehicle data is input into the target scene discrimination model, and the scene where the target vehicle is located is output.

[0058] Here, the preset scene discrimination model can be trained in advance to obtain the target scene discrimination model.

[0059] Specifically, training data may be obtained, which may include vehicle data with scene labels, including environmental data and driving data. A neural network is trained using the training data to obtain a trained target scene discrimination model.

[0060] In some embodiments, the target scene discrimination model can be a combination model of a convolutional network and a fully connected network, or a combination model of a long short-term memory network and a fully connected network.

[0061] In step 110 , training data is obtained, the input information is vehicle environment information and state information, and the output is the scene in which the vehicle is located when the training data is obtained.

[0062] In some embodiments, to more accurately determine the scene in which the target vehicle is located, the above S120 may include:

[0063] Input the first vehicle data into the convolutional network and output a first feature;

[0064] The first feature and the second vehicle data are input into the fully connected network, and the scene where the target vehicle is located is output.

[0065] Here, the first vehicle data may include lane line data, obstacle data, first speed, first position, target speed and target position. The second vehicle data may include target environment data and target driving data other than the first vehicle data and driving data.

[0066] During the training process of the combined convolutional network and fully connected network model, the convolutional network processes the spatial input. The input is divided into multiple layers: one layer is an abstract grid of lanes and curbs, with 1s and 0s representing the presence or absence of lane lines at the corresponding location; one layer is an obstacle layer, also with 1s and 0s representing the presence or absence of obstacles at the corresponding location; and one layer is a vehicle information layer, with values ​​representing the speeds of nearby vehicles and the target vehicle. The first feature output by the convolutional network is combined with other inputs as the input to the fully connected network. The final layer of the fully connected network is a normalized exponential function layer (softmax layer), which outputs the probability of each scene. This can be compared with the standard output of the one-hot encoded training data, and the model is trained via gradient descent. The combined model (i.e., the preset scene discrimination model) can then be continuously trained until the output meets the training stopping condition, resulting in the trained target scene discrimination model. The training stopping condition can be that the output meets the user's preset requirements, which is not limited here.

[0067] In this way, through the above process, the scene in which the target vehicle is located can be determined more accurately.

[0068] In some embodiments, in order to more quickly determine the scene in which the target vehicle is located, the above S120 may include:

[0069] Input the lane centerline point sequence into the long short-term memory network and output the second feature;

[0070] The second feature and the third vehicle data are input into the fully connected network, and the scene where the target vehicle is located is output.

[0071] Here, the third vehicle data may include other environmental data and driving data except for the point sequence of the lane center line in the target environmental data and the target driving data.

[0072] In the process of training the combined model of the long short-term memory network and the fully connected network, the long short-term memory network can be used to process the point sequence of the lane centerline, etc. After extracting the second feature, the second feature and other non-sequence inputs are input into the fully connected network together. The last layer of the fully connected network is a normalized exponential function layer (softmax layer), so the probability of each scene can be output, which can be compared with the standard output of the training data after one-hot encoding, and the model can be trained by gradient descent. Then, the combined model (that is, the preset scene discrimination model) can be continuously trained until the output result meets the training stopping condition, thereby obtaining the trained target scene discrimination model. The training stopping condition can be that the output result meets the user's preset requirements, which is not limited here.

[0073] In this way, the use of long short-term memory networks can significantly reduce the amount of computation, thereby determining the scene in which the target vehicle is located more quickly.

[0074] Regarding S130, if the scene in which the target vehicle is located is not the target scene, the second reference line obtained by smoothing the lane center line can be directly used as the target reference line without adjustment.

[0075] If the scene in which the target vehicle is located is a target scene, multiple target control points may be determined according to a control point determination method corresponding to the target scene.

[0076] In some implementations, the target scenario may include: a lane merging scenario, a left turn scenario, a right turn scenario, and a U-turn scenario.

[0077] Specifically, lane-merging scenarios may include, but are not limited to, congested and uncongested scenarios. Left-turn scenarios may include, but are not limited to, left-turn red light scenarios and left-turn green light scenarios, which can be further subdivided into left-turn green light scenarios with a waiting area and left-turn green light scenarios without a waiting area. Right-turn scenarios may include, but are not limited to, right-turn red light scenarios and right-turn green light scenarios. Other scenarios may also be set as target scenarios based on actual circumstances, and are not limited here.

[0078] In some embodiments, the multiple target control points corresponding to the lane merging scene can be as follows: Figure 2 As shown, control point 202 is the end point of the front lane segment of the merging lane segment; control point 201 is the point at the first preset distance behind the vehicle from control point 202 along the front lane segment; control point 204 is obtained by interpolating a point at a first longitudinal distance (the first longitudinal distance refers to the distance from the projection of the point in the front lane segment to the control point 202) from control point 202 on the subsequent lane segment of the merging lane segment, and the specific value of the first longitudinal distance is calculated by the transverse distance (the distance from the point to the extension line of the front lane segment); control point 203 is the second preset distance behind the vehicle from control point 204 along the subsequent lane segment.

[0079] The multiple target control points corresponding to the left turn scene can be as follows Figure 3 As shown, control point 301 is a point on the extension line of the first lane segment before the stop line of the starting intersection, which is the third preset distance away from the stop line, and the third preset distance is related to the curb length; control point 302 is a point on the extension line of the first lane segment before the stop line of the starting intersection, which is the fourth preset distance away from control point 301; control point 303 is the intersection of the extension line of the center line of the first lane at the stop line of the starting and target intersections; control point 304 is on the extension line of the first lane after the stop line of the target intersection, which is the fifth preset distance away from the stop line; control point 305 is on the extension line of the first lane after the stop line of the target intersection, which is the sixth preset distance away from the stop line.

[0080] The multiple target control points corresponding to the right turn scene can be as follows Figure 4 As shown, control point 401 is a point on the extension line of the first lane segment before the stop line of the starting intersection, which is the seventh preset distance from the stop line, and the seventh preset distance is related to the curb length; control point 402 is a point on the extension line of the first lane segment before the stop line of the starting intersection, which is the eighth preset distance from control point 401; control point 403 is the intersection of the extension line of the center line of the first lane at the stop line of the starting and target intersections; control point 404 is on the extension line of the first lane after the stop line of the target intersection, which is the ninth preset distance from the stop line; control point 405 is on the extension line of the first lane after the stop line of the target intersection, which is the tenth preset distance from the stop line.

[0081] The multiple target control points corresponding to the U-turn scenario can be Figure 5As shown, control point 501 is a point on the second reference line that is the eleventh preset distance from the stop line at the intersection, and the eleventh preset distance is related to the length of the curb; control point 506 is a point on the second reference line that is the twelfth preset distance from the stop line at the intersection, and the twelfth preset distance is related to the length of the curb; control point 502 is a point on the second reference line that is the thirteenth preset distance from control point 501; control point 505 is a point on the second reference line that is the fourteenth preset distance from control point 506; control point 503 is on a parallel line to control point 501 and can be adjusted left and right; control point 504 is on a parallel line to control point 505 and can be adjusted left and right.

[0082] In the target scene, it is often easy to find that the second reference line obtained by directly smoothing the lane centerline is unreasonable, so the second reference line needs to be optimized.

[0083] In S140 , a curve may be used to fit a plurality of target control points to obtain a first reference line. For example, a Bezier curve may be used for fitting.

[0084] Regarding S150 , the second reference line may be obtained by smoothing the lane centerline. The second reference line may also refer to a sequence that can form or create a reference line, such as a point sequence, a curve sequence, a lane sequence, etc.

[0085] In some implementations, to obtain a reference line that is closer to human driving habits, the above S150 may include:

[0086] determining a target lane segment corresponding to the first reference line and multiple lane segments corresponding to the second reference line;

[0087] The reference line area corresponding to the target lane segment in the second reference line is replaced by the first reference line.

[0088] Here, the length of the first reference line may be shorter than that of the second reference line. The first reference line may be a local reference line, and the second reference line may be optimized by replacing a part of the second reference line with the first reference line to obtain a target reference line.

[0089] For example, in a lane merging scenario, the unoptimized second reference line and the optimized target reference line can be as follows: Figure 2 As shown, the solid line is the second reference line and the dotted line is the target reference line.

[0090] In the left-turn scenario, the unoptimized second reference line and the optimized target reference line can be Figure 3 As shown, the unmarked solid line is the second reference line, and the dotted line is the target reference line.

[0091] In the right turn scenario, the unoptimized second reference line and the optimized target reference line can be Figure 4 As shown, the unmarked solid line is the second reference line, and the dotted line is the target reference line.

[0092] In the U-turn scenario, the unoptimized second reference line and the optimized target reference line can be Figure 5 As shown, the unmarked solid line is the second reference line, and the dotted line is the target reference line.

[0093] In this way, by optimizing the second reference line through the first reference line, a target reference line that is closer to human driving habits can be obtained, making autonomous driving more intelligent.

[0094] The reference line determination method provided in the embodiment of the present application can make the driving trajectory of an autonomous driving vehicle closer to the driving trajectory of human driving, avoiding sudden changes in steering angle and curvature that cause vehicle instability.

[0095] Based on the same inventive concept, the present application embodiment also provides a reference line determination device. Figure 6 The reference line determination device provided in the embodiment of the present application is described in detail.

[0096] Figure 6 A schematic structural diagram of a reference line determination device provided by an embodiment of the present application is shown.

[0097] like Figure 6 As shown, the reference line determination device may include:

[0098] An acquisition module 601 is configured to acquire target vehicle data corresponding to a target vehicle, wherein the target vehicle data includes target environment data and target driving data;

[0099] A first determining module 602 is configured to determine a scene in which a target vehicle is located based on target vehicle data;

[0100] A second determining module 603 is configured to determine a plurality of target control points according to a control point determination method corresponding to the target scene when the scene is a target scene;

[0101] A creation module 604 is configured to create a first reference line based on a plurality of target control points;

[0102] The adjustment module 605 is configured to adjust the second reference line based on the first reference line to obtain a target reference line, where the second reference line is obtained by smoothing the lane center line.

[0103] In this way, the target vehicle data corresponding to the target vehicle can be obtained, and the target vehicle data includes target environment data and target driving data. Then, the scene in which the target vehicle is located can be determined based on the target vehicle data. In the case where the scene is a target scene, multiple target control points are determined according to the control point determination method corresponding to the target scene. Then, a first reference line is created based on the multiple target control points. Then, based on the first reference line, a second reference line is adjusted to obtain a target reference line. The second reference line is obtained by smoothing the center line of the lane. In other words, the first reference line can be determined based on the specific scene in which the target vehicle is located, and the second reference line obtained by directly smoothing the center line of the lane can be adjusted based on the first reference line to obtain the target reference line. In this way, the target reference line can be made more suitable for the scene in which the target vehicle is currently located, and the driving trajectory generated based on the target reference line is more consistent with the driving trajectory of a human-driven vehicle, thereby improving the intelligence of the autonomous driving vehicle.

[0104] In some embodiments, to more accurately determine the scene in which the target vehicle is located, the target environment data includes traffic signal data, road data, obstacle data, and a first speed and a first position of a first vehicle whose distance from the target vehicle does not exceed a distance threshold; the road data includes a lane position, lane line data, a point sequence of a lane centerline, and lane turning attributes;

[0105] The target driving data includes the target speed, target acceleration and target position of the target vehicle.

[0106] In some implementations, to more accurately determine the scene in which the target vehicle is located, the first determination module 602 may include:

[0107] A first input submodule is configured to input first vehicle data into a convolutional network and output a first feature, wherein the first vehicle data includes lane line data, obstacle data, a first speed, a first position, a target speed, and a target position;

[0108] The second input submodule is used to input the first feature and the second vehicle data into the fully connected network, and output the scene in which the target vehicle is located, where the second vehicle data includes other environmental data and driving data in the target environmental data and the target driving data except the first vehicle data.

[0109] In some implementations, in order to more quickly determine the scene in which the target vehicle is located, the first determination module 602 may include:

[0110] The third input submodule is used to input the lane centerline point sequence into the long short-term memory network and output the second feature;

[0111] The fourth input submodule is used to input the second feature and the third vehicle data into the fully connected network, and output the scene in which the target vehicle is located. The third vehicle data includes other environmental data and driving data in the target environment data and the target driving data except the point sequence of the lane centerline.

[0112] In some implementations, the target scenarios include: a lane merging scenario, a left turn scenario, a right turn scenario, and a U-turn scenario.

[0113] In some implementations, to obtain a reference line that is closer to human driving habits, the adjustment module 605 may include:

[0114] A determination submodule, configured to determine a target lane segment corresponding to the first reference line and multiple lane segments corresponding to the second reference line;

[0115] The replacement submodule is used to replace the reference line area corresponding to the target lane segment in the second reference line with the first reference line, where the length of the first reference line is shorter than the length of the second reference line.

[0116] Figure 7 A schematic structural diagram of an electronic device provided by an embodiment of the present application is shown.

[0117] like Figure 7 As shown in FIG, the electronic device 7 is a block diagram of an exemplary hardware architecture of an electronic device capable of implementing the reference line determination method and reference line determination apparatus according to the embodiments of the present application. The electronic device may refer to the electronic device in the embodiments of the present application.

[0118] The electronic device 7 may include a processor 701 and a memory 702 storing computer program instructions.

[0119] Specifically, the processor 701 may include a central processing unit (CPU), or an application-specific integrated circuit (ASIC), or may be configured to implement one or more integrated circuits of the embodiments of the present application.

[0120] The memory 702 may include a large-capacity memory for data or instructions. By way of example and not limitation, the memory 702 may include a hard disk drive (HDD), a floppy disk drive, flash memory, an optical disk, a magneto-optical disk, a magnetic tape, or a universal serial bus (USB) drive, or a combination of two or more of these. Where appropriate, the memory 702 may include removable or non-removable (or fixed) media. Where appropriate, the memory 702 may be internal or external to the integrated gateway disaster recovery device. In certain embodiments, the memory 702 is a non-volatile solid-state memory. In certain embodiments, the memory 702 may include a read-only memory (ROM), a random access memory (RAM), a magnetic disk storage media device, an optical storage media device, a flash memory device, an electrical, optical, or other physical / tangible memory storage device. Therefore, generally, the memory 702 includes one or more tangible (non-transitory) computer-readable storage media (e.g., a memory device) encoded with software including computer-executable instructions, and when the software is executed (e.g., by one or more processors), it is operable to perform the operations described with reference to the method according to one aspect of the present application.

[0121] The processor 701 reads and executes computer program instructions stored in the memory 702 to implement any one of the reference line determination methods in the above embodiments.

[0122] In one example, the electronic device may further include a communication interface 703 and a bus 704. Figure 7 As shown, the processor 701 , the memory 702 , and the communication interface 703 are connected via a bus 704 and communicate with each other.

[0123] The communication interface 703 is mainly used to implement communication between various modules, devices, units and / or equipment in the embodiments of the present application.

[0124] Bus 704 comprises hardware, software or both, couples the parts of electronic equipment to each other.For example, and not limitation, bus can comprise accelerated graphics port (AGP) or other graphics bus, enhanced industry standard architecture (EISA) bus, front side bus (FSB), hypertransport (HT) interconnection, industry standard architecture (ISA) bus, infinite bandwidth interconnection, low pin count (LPC) bus, memory bus, micro channel architecture (MCA) bus, peripheral component interconnection (PCI) bus, PCI-Express (PCI-X) bus, serial advanced technology attachment (SATA) bus, video electronics standard association local (VLB) bus or other suitable bus or two or more of these combinations.In suitable cases, bus 704 can comprise one or more buses.Although the present application embodiment describes and shows specific bus, the application considers any suitable bus or interconnection.

[0125] The electronic device can execute the reference line determination method in the embodiment of the present application, thereby realizing the combination Figures 1 to 6 The reference line determination method and device described.

[0126] In addition, in conjunction with the reference line determination method in the above embodiments, embodiments of the present application may provide a computer storage medium for implementation. The computer storage medium stores computer program instructions; when the computer program instructions are executed by a processor, any of the reference line determination methods in the above embodiments is implemented.

[0127] It should be understood that the present application is not limited to the specific configurations and processes described above and illustrated in the figures. For the sake of brevity, a detailed description of known methods is omitted here. In the above embodiments, several specific steps are described and illustrated as examples. However, the method process of the present application is not limited to the specific steps described and illustrated. Those skilled in the art can make various changes, modifications, and additions, or change the order of the steps after understanding the spirit of the present application.

[0128] The functional blocks shown in the above-described block diagram can be implemented as hardware, software, firmware or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of the present application are programs or code segments that are used to perform the required tasks. The program or code segment can be stored in a machine-readable medium, or transmitted on a transmission medium or a communication link by a data signal carried in a carrier wave. "Machine-readable medium" can include any medium that can store or transmit information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROMs, flash memories, erasable ROMs (EROMs), floppy disks, CD-ROMs, optical disks, hard disks, optical fiber media, radio frequency (RF) links, etc. The code segment can be downloaded via a computer network such as the Internet, an intranet, etc.

[0129] It should also be noted that the exemplary embodiments mentioned in this application describe some methods or systems based on a series of steps or devices. However, this application is not limited to the order of the above steps. In other words, the steps can be performed in the order mentioned in the embodiments, or in a different order, or several steps can be performed simultaneously.

[0130] Aspects of the present application have been described above with reference to the flowcharts and / or block diagrams of the methods, devices (systems) and computer program products according to the embodiments of the present application. It should be understood that each box in the flowchart and / or block diagram and the combination of each box in the flowchart and / or block diagram can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer or other programmable data processing device to produce a machine so that these instructions executed via the processor of the computer or other programmable data processing device enable the implementation of the function / action specified in one or more boxes of the flowchart and / or block diagram. This processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special application processor or a field programmable logic circuit. It is also understood that each box in the block diagram and / or the flowchart and the combination of the boxes in the block diagram and / or the flowchart can also be implemented by the dedicated hardware that performs the specified function or action, or can be implemented by the combination of dedicated hardware and computer instructions.

[0131] The above description is only a specific embodiment of the present application. Those skilled in the art will clearly understand that for the convenience and brevity of description, the specific working processes of the systems, modules and units described above can refer to the corresponding processes in the aforementioned method embodiments, and will not be repeated here. It should be understood that the scope of protection of the present application is not limited thereto. Any person skilled in the art can easily think of various equivalent modifications or replacements within the technical scope disclosed in the present application, and these modifications or replacements should be included in the scope of protection of the present application.

Claims

1. A reference line determination method, characterized in that: include: Acquire target vehicle data corresponding to the target vehicle, wherein the target vehicle data includes target environment data and target driving data; Determining the scene in which the target vehicle is located according to the target vehicle data; In a case where the scene is a target scene, determining a plurality of target control points according to a control point determination method corresponding to the target scene; creating a first reference line based on the plurality of target control points; Adjusting a second reference line based on the first reference line to obtain a target reference line, where the second reference line is obtained by smoothing the lane centerline; The target environment data includes traffic signal data, road data, obstacle data, and a first speed and a first position of a first vehicle whose distance from the target vehicle does not exceed a distance threshold, and the road data includes a lane position, lane line data, a point sequence of a lane centerline, and a lane turning attribute; The target driving data includes a target speed, a target acceleration and a target position of the target vehicle; The determining the scene in which the target vehicle is located according to the target vehicle data includes: Inputting first vehicle data into a convolutional network and outputting a first feature, wherein the first vehicle data includes the lane line data, the obstacle data, the first speed, the first position, the target speed, and the target position; Inputting the first feature and the second vehicle data into a fully connected network, and outputting a scene in which the target vehicle is located, wherein the second vehicle data includes other environmental data and driving data in the target environmental data and the target driving data except the first vehicle data; The adjusting the second reference line based on the first reference line to obtain a target reference line includes: determining a target lane segment corresponding to the first reference line and a plurality of lane segments corresponding to the second reference line; A reference line region of the second reference line corresponding to the target lane segment is replaced with the first reference line, where the length of the first reference line is shorter than that of the second reference line.

2. The method according to claim 1, characterized in that The determining the scene in which the target vehicle is located according to the target vehicle data includes: Inputting the lane centerline point sequence into a long short-term memory network and outputting a second feature; The second feature and the third vehicle data are input into a fully connected network, and the scene in which the target vehicle is located is outputted, wherein the third vehicle data includes other environmental data and driving data in the target environmental data and the target driving data except the point sequence of the lane centerline.

3. The method according to claim 1, characterized in that The target scenarios include: lane merging scenario, left turn scenario, right turn scenario and U-turn scenario.

4. A reference line determination device, characterized in that: The device comprises: An acquisition module is used to acquire target vehicle data corresponding to a target vehicle, wherein the target vehicle data includes target environment data and target driving data; A first determining module, configured to determine a scene in which the target vehicle is located based on the target vehicle data; a second determining module, configured to determine a plurality of target control points according to a control point determination method corresponding to the target scene when the scene is a target scene; A creating module, configured to create a first reference line based on the plurality of target control points; an adjustment module, configured to adjust a second reference line based on the first reference line to obtain a target reference line, wherein the second reference line is obtained by smoothing the lane centerline; The target environment data includes traffic signal data, road data, obstacle data, and a first speed and a first position of a first vehicle whose distance from the target vehicle does not exceed a distance threshold, and the road data includes a lane position, lane line data, a point sequence of a lane centerline, and a lane turning attribute; The target driving data includes a target speed, a target acceleration and a target position of the target vehicle; The determining the scene in which the target vehicle is located according to the target vehicle data includes: Inputting first vehicle data into a convolutional network and outputting a first feature, wherein the first vehicle data includes the lane line data, the obstacle data, the first speed, the first position, the target speed, and the target position; Inputting the first feature and the second vehicle data into a fully connected network, and outputting a scene in which the target vehicle is located, wherein the second vehicle data includes other environmental data and driving data in the target environmental data and the target driving data except the first vehicle data; The adjusting the second reference line based on the first reference line to obtain a target reference line includes: determining a target lane segment corresponding to the first reference line and a plurality of lane segments corresponding to the second reference line; A reference line region of the second reference line corresponding to the target lane segment is replaced with the first reference line, where the length of the first reference line is shorter than that of the second reference line.

5. An electronic device, characterized in that: The device includes: a processor and a memory storing computer program instructions; When the processor executes the computer program instructions, the reference line determination method according to any one of claims 1 to 3 is implemented.

6. A computer storage medium, characterized in that The computer storage medium stores computer program instructions, which, when executed by a processor, implement the reference line determination method according to any one of claims 1 to 3.

7. A computer program product, characterized in that When the instructions in the computer program product are executed by a processor of an electronic device, the electronic device is caused to execute the reference line determination method according to any one of claims 1 to 3.

Images