Reference line generation method and device, vehicle and storage medium

Abstract

The invention relates to a reference line generation method and device, a vehicle and a storage medium. The reference line generation method comprises the steps that turning information of a turning intersection is determined according to a navigation path of a vehicle and / or road semantic information detected by the vehicle, and the turning information comprises a turning starting point and a turning ending point; and generating a turning path reference line of the turning intersection according to the turning starting point and the turning ending point. The method can improve the safety and stability of automatic driving.

Description

Technical Field

[0001] The present disclosure relates to the technical field of vehicles, and particularly to a reference line generation method, apparatus, vehicle, and storage medium. Background Art

[0002] With the development of vehicle technology, autonomous driving technology has also developed. Autonomous driving technology is a technology that enables a vehicle to autonomously perform environmental perception, path planning, and vehicle control through technologies such as artificial intelligence and machine learning. Autonomous driving technology can be applied to various traffic scenarios, including urban roads, highways, parking lots, etc., to achieve more intelligent, safe, and convenient transportation. Summary of the Invention

[0003] To overcome the problems existing in the related art, the present disclosure provides a reference line generation method, apparatus, vehicle, and storage medium, which are used to improve the safety and stability of autonomous driving.

[0004] According to a first aspect of an embodiment of the present disclosure, a reference line generation method is provided, including: determining turning information of a turning intersection according to a navigation path of a vehicle and / or road semantic information detected by the vehicle, where the turning information includes a turning start point and a turning end point; generating a turning path reference line of the turning intersection according to the turning start point and the turning end point.

[0005] Optionally, the determining turning information of a turning intersection according to a navigation path of a vehicle and / or road semantic information detected by the vehicle includes: obtaining path point orientation information of the navigation path, where the path point orientation information includes orientations respectively corresponding to a plurality of navigation path points; determining a first navigation path point and a second navigation path point according to the path point orientation information, where the orientation of the first navigation path point meets a preset turning start point condition, and the orientation of the second navigation path point meets a preset turning end point condition; determining the turning start point according to the first navigation path point, and determining the turning end point according to the second navigation path point.

[0006] Optionally, the road semantic information includes: zebra crossing information, and the determining turning information of a turning intersection according to a navigation path of a vehicle and / or road semantic information detected by the vehicle includes: determining the turning start point and the turning end point according to the zebra crossing information and a preset positional relationship, where the preset positional relationship is used to characterize the positional relationships between the turning start point and the turning end point and the zebra crossing respectively.

[0007] Optionally, the zebra crossing information includes first zebra crossing information and second zebra crossing information corresponding to the turning direction of the turning intersection. Determining the turning start point and the turning end point according to the zebra crossing information and a preset positional relationship includes: determining the turning start point according to the first zebra crossing information and a preset first positional relationship, where the first positional relationship is used to characterize the position of the turning start point relative to the first zebra crossing; determining the turning end point according to the second zebra crossing information and a preset second positional relationship, where the second positional relationship is used to characterize the position of the turning end point relative to the second zebra crossing.

[0008] Optionally, determining the turning information of the turning intersection according to the navigation path of the vehicle and / or the road semantic information detected by the vehicle includes: determining first turning information according to the navigation path, where the first turning information includes a first turning start point and a first turning end point; determining second turning information according to the road semantic information, where the second turning information includes a second turning start point and a second turning end point; determining target turning information according to the first turning information, the second turning information and a straight path reference line, where the target turning information includes a target turning start point and a target turning end point.

[0009] Optionally, the straight path reference line includes a straight path reference line corresponding to the turning start point and a straight path reference line corresponding to the turning end point. Determining the target turning information according to the first turning information, the second turning information and the straight path reference line includes: projecting the first turning start point and the first turning end point onto the corresponding straight path reference lines respectively to determine the projected first turning start point and the projected first turning end point; projecting the second turning start point and the second turning end point onto the corresponding straight path reference lines respectively to determine the projected second turning start point and the projected second turning end point; determining the target turning start point according to the first turning start point, the second turning start point, the projected first turning start point and the projected second turning start point; determining the target turning end point according to the first turning end point, the second turning end point, the projected first turning end point and the projected second turning end point.

[0010] Optionally, determining the target turning starting point according to the first turning starting point, the second turning starting point, the projected first turning starting point, and the projected second turning starting point includes: determining the target turning starting point according to the first turning starting point, the weight corresponding to the navigation path, the second turning starting point, the weight corresponding to the road semantic information, the projected first turning starting point, the projected second turning starting point, and the weight corresponding to the straight path reference line; determining the target turning ending point according to the first turning ending point, the second turning ending point, the projected first turning ending point, and the projected second turning ending point includes: determining the target turning ending point according to the first turning ending point, the weight corresponding to the navigation path, the second turning ending point, the weight corresponding to the road semantic information, the projected first turning ending point, the projected second turning ending point, and the weight corresponding to the straight path reference line.

[0011] Optionally, the reference line generation method further includes: projecting the target turning starting point onto the corresponding straight path reference line to determine the projected target turning starting point; determining the final turning starting point according to the projected target turning starting point.

[0012] Optionally, generating the turning path reference line of the turning intersection according to the turning starting point and the turning ending point includes: generating the turning path reference line of the turning intersection according to the turning starting point, the turning ending point, and a preset curve generation algorithm.

[0013] Optionally, the preset curve generation algorithm is a second-order Bezier curve algorithm, and the control point of the second-order Bezier curve algorithm is the intersection of the turning starting point straight line and the turning ending point straight line. The turning starting point straight line is determined according to the turning starting point and the first orientation, and the turning ending point straight line is determined according to the turning ending point and the second orientation.

[0014] Optionally, the reference line generation method further includes: in response to the vehicle detecting new road semantic information, determining a new turning ending point according to the new road semantic information; adjusting the turning path reference line according to the new turning ending point to obtain an adjusted turning path reference line.

[0015] Optionally, adjusting the turning path reference line according to the new turning ending point to obtain an adjusted turning path reference line includes: determining the adjusted turning ending point according to the new turning ending point, the weight corresponding to the new turning ending point, the turning ending point, and the weight corresponding to the turning ending point; determining the adjusted turning path reference line according to the turning starting point and the adjusted turning ending point.

[0016] Optionally, the weight corresponding to the new turning ending point is less than the weight corresponding to the turning ending point.

[0017] According to a second aspect of the embodiments of the present disclosure, a reference line generation device is provided, including: a determination module configured to determine turning information of a turning intersection according to a navigation path of a vehicle and / or road semantic information detected by the vehicle, where the turning information includes a turning start point and a turning end point; a generation module configured to generate a turning path reference line of the turning intersection according to the turning start point and the turning end point.

[0018] According to a third aspect of the embodiments of the present disclosure, a computer-readable storage medium is provided, on which computer program instructions are stored, and when the program instructions are executed by a processor, the steps of the reference line generation method provided in the first aspect of the present disclosure are implemented.

[0019] According to a fourth aspect of the embodiments of the present disclosure, a vehicle is provided, including: a processor; a memory for storing processor-executable instructions; wherein, the processor is configured to execute the steps of the reference line generation method provided in the first aspect of the present disclosure.

[0020] The technical solutions provided by the embodiments of the present disclosure may include the following beneficial effects: By the navigation path of the vehicle and / or the road semantic information detected by the vehicle, the turning information of the turning intersection is determined, and by using the turning information, a turning path reference line can be generated, and the turning path reference line can be used to implement the planning of the turning path. Since the generation of the turning path reference line does not depend on a high-precision map; then, for a vehicle, the turning path reference line can be generated even without a high-precision map. Thus, the technical solution can be applied to more autonomous driving scenarios or more traffic scenarios, improving the safety and stability of autonomous driving.

[0021] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The accompanying drawings herein are incorporated into the specification and constitute a part of the specification, showing embodiments consistent with the present disclosure, and are used together with the specification to explain the principles of the present disclosure.

[0023] Figure 1 is a flowchart of a reference line generation method shown according to an exemplary embodiment.

[0024] Figure 2 is a schematic diagram of a turning intersection shown according to an exemplary embodiment.

[0025] Figure 3 is a schematic diagram of a first type of turning information shown according to an exemplary embodiment.

[0026] Figure 4It is a schematic diagram of the second turning information shown according to an exemplary embodiment.

[0027] Figure 5 It is a schematic diagram of the third turning information shown according to an exemplary embodiment.

[0028] Figure 6 It is a schematic diagram of the turning end point adjustment shown according to an exemplary embodiment.

[0029] Figure 7 It is a block diagram of a reference line generation device shown according to an exemplary embodiment.

[0030] Figure 8 It is a schematic block diagram of the functions of a vehicle shown according to an exemplary embodiment. Detailed implementation manners

[0031] Here, the exemplary embodiments will be described in detail, and the examples are shown in the drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementation manners described in the following exemplary embodiments do not represent all implementation manners consistent with the present disclosure. On the contrary, they are merely examples of devices and methods consistent with some aspects of the present disclosure as detailed in the appended claims.

[0032] It should be noted that all actions of obtaining signals, information, or data in this application are carried out on the premise of complying with the corresponding data protection regulations and policies of the country where it is located and obtaining the authorization given by the owner of the corresponding device.

[0033] The technical solutions of the embodiments of the present disclosure can be applied to various autonomous driving scenarios. In these autonomous driving scenarios, through autonomous driving technology, the vehicle can autonomously perform environmental perception, path planning, and vehicle control.

[0034] For autonomous driving technology, it can be divided into multiple levels: assisted driving, partial autonomous driving, conditional autonomous driving, highly autonomous driving, and fully autonomous driving. For different levels, the degree of automation is different. Relatively speaking, the degree of automation of assisted driving is the lowest, and the degree of automation of fully autonomous driving is the highest.

[0035] In autonomous driving scenarios, it is relatively dependent on high-precision maps. A high-precision map is a high-precision map for autonomous driving, which contains map elements such as road shapes, road markings, traffic signs, and obstacles; the map accuracy can reach the centimeter level. A high-precision map is one of the core technologies of autonomous driving vehicles and can help autonomous driving vehicles achieve more accurate, safe, and efficient autonomous driving.

[0036] However, in some special scenarios, there is no high-precision map. In the case of no high-precision map, it is difficult to perform the turning operation at a turning intersection, thus unable to ensure the safety and stability of autonomous driving.

[0037] In the related art, in the case of no high-precision map, no suitable solution has been found for how to perform the turning operation at a turning intersection to ensure the safety and stability of autonomous driving.

[0038] Based on this, the embodiments of the present disclosure provide a technical solution. By using various available information that can be obtained in the scenario without a high-precision map, a turning path reference line that can be used to guide the turning at a turning intersection is generated to improve the safety and stability of autonomous driving.

[0039] Figure 1 is a flowchart of a reference line generation method shown according to an exemplary embodiment. As Figure 1 shown, the reference line generation method is used in a vehicle and includes the following steps:

[0040] Step S11: Determine the turning information of the turning intersection according to the navigation path of the vehicle and / or the road semantic information detected by the vehicle; the turning information includes a turning start point and a turning end point.

[0041] Step S12: Generate a turning path reference line for the turning intersection according to the turning start point and the turning end point.

[0042] In step S11, the turning intersection may be a left-turn intersection or a right-turn intersection. The turning intersection may be the turning intersection that the vehicle is about to reach. That is, before the vehicle is about to reach the turning intersection, it is necessary to generate a turning path reference line for the turning intersection to plan a turning path based on the turning path reference line.

[0043] Please refer to Figure 2 , which is a schematic diagram of a turning intersection shown according to an exemplary embodiment. As Figure 2 shown, the turning intersection is a cross-turn intersection. If the vehicle is on the left lane, it may perform a left-turn operation; if the vehicle is on the right lane, it may perform a right-turn operation. Figure 2 Taking the right-turn operation as an example in , based on the center line of the lane, a turning reference line for the intersection can be generated; through the turning reference line for the intersection, a turning path can be generated.

[0044] In step S11, two types of information may be involved. One is the navigation path of the vehicle, and the other is the road semantic information detected by the vehicle. These two types of information can be combined to determine the turning information, or these two types of information can be used separately to determine the turning information.

[0045] In different traffic scenarios, different methods can be used to determine turning information. For example, if only one of the pieces of information is used to determine the turning information, the resulting turning path reference line may not be very accurate. Therefore, in traffic scenarios where the accuracy requirement is not very high, only one of the pieces of information can be used to determine the turning information. For example: traffic scenarios with relatively simple road conditions. In traffic scenarios where a higher accuracy is required, two pieces of information can be combined to determine the turning information; for example: traffic scenarios with relatively complex road conditions.

[0046] In some embodiments, the navigation path of the vehicle can be the path provided by the navigation map. This navigation map is not a high-precision map but an ordinary map, and the map information it contains is less compared to a high-precision map. For example, this navigation map can be the map provided by the vehicle navigation system or the map provided by the navigation software; thus, the navigation path can be the navigation path provided by the vehicle navigation system or the navigation path provided by the navigation software.

[0047] In some embodiments, the navigation path includes multiple navigation path points, and these multiple navigation path points form the navigation path.

[0048] As an alternative implementation, step S11 includes: obtaining the path point orientation information of the navigation path, where the path point orientation information includes the orientations corresponding to multiple navigation path points respectively; according to the path point orientation information, determining a first navigation path point and a second navigation path point, where the orientation of the first navigation path point meets the preset turning start point condition, and the orientation of the second navigation path point meets the preset turning end point condition; determining the turning start point according to the first navigation path point and determining the turning end point according to the second navigation path point.

[0049] In some embodiments, the path point orientation information can be the information inherent in the navigation map or the information calculated based on the position information of each navigation point in the navigation map.

[0050] In this implementation, the orientation corresponding to each navigation path point can be understood as the angle between the vector formed by this navigation path point and the next navigation path point and the base vector, and this base vector can be a vector representing the base direction, such as: a horizontal direction vector or a vertical direction vector, etc.

[0051] Therefore, the orientation corresponding to each navigation path point can also be understood as the orientation of the navigation path at the position of this navigation path point; that is, the navigation path has different orientations at different positions.

[0052] In some embodiments, the preset turning start conditions may be: the difference from the orientation of the previous navigation path point is greater than a preset value, and / or the difference from the orientation of the next navigation path point is greater than a preset value; the preset turning end conditions may be: the difference from the orientation of the previous navigation path point is greater than a preset value, and the difference from the orientation of the next navigation path point is less than a preset value. Wherein, the preset value can be configured according to different traffic scenarios.

[0053] Thus, the first navigation path point may be the point where the orientation starts to change significantly for the first time, and the second navigation path point may be the point where the orientation changes significantly for the last time.

[0054] In some embodiments, the first navigation path point may be determined as the turning start point, or a neighboring navigation path point (such as the previous or the next one) of the first navigation path point may be determined as the turning start point. And, the second navigation path point may be determined as the turning end point, or a neighboring navigation path point of the second navigation path point may be determined as the turning end point.

[0055] In this implementation, since the orientation of the turning intersection on the navigation path is different from that of other positions, the turning start point and the turning end point can be simply determined through the orientation information of each path point on the navigation path.

[0056] In some embodiments, in combination with the path point orientation information, it can also be determined whether the vehicle is about to reach the turning intersection, and then it can be determined whether a turning path reference line needs to be generated.

[0057] Therefore, as an alternative implementation, when the first navigation path point where the orientation starts to change significantly is detected, step S11 is executed. Wherein, the first navigation path point where the orientation starts to change significantly may be the aforementioned first navigation path point.

[0058] Thus, if the prerequisite for executing step S11 is to detect the first navigation path point, then when determining the turning information based on the navigation path, it is not necessary to repeatedly determine the first navigation path point.

[0059] In some embodiments, detecting whether the orientation of the navigation path point starts to change significantly may be an operation that is continuously repeated. Thus, the planning of the turning path can start before the vehicle reaches the turning intersection.

[0060] In the embodiments of the present disclosure, the accuracy of the navigation path point may be at the road level, and there may be a certain degree of error. Therefore, in scenarios where the accuracy requirement is not high, this implementation can be used to determine the turning information.

[0061] Please refer to Figure 3, which is a schematic diagram of the first turning information shown according to an exemplary embodiment. In Figure 3 , a plurality of navigation points (i.e., navigation path points) are shown, as well as a turning start point and a turning end point determined according to the navigation path. Moreover, these navigation points all have orientation information, and the direction indicated by the arrow is the orientation.

[0062] In the embodiments of the present disclosure, road semantic information refers to multi-level and rich-dimensional information that can enable an autonomous vehicle to better understand driving rules, perceive road traffic conditions, and plan driving routes. Such information can not only include geometric information such as the shape, size, traffic signs, and lane lines of the road, but also include dynamic information such as traffic rules and traffic flow, as well as static information such as road surface conditions and obstacles.

[0063] This road semantic information can be obtained through the vehicle's perception technology. Vehicle perception technology refers to the use of technologies such as sensors and computer vision to achieve the perception and understanding of the information in the vehicle's surrounding environment. Vehicle perception technology is an important part of autonomous driving technology. It can help an autonomous vehicle perceive the surrounding environment, including information such as road boundaries, vehicles, and pedestrians, so as to provide a basis for the decision-making and control of the autonomous vehicle. Vehicle perception technology can involve the following aspects:

[0064] Sensor technology: including laser rangefinders, video cameras, vehicle-mounted radars, speed and acceleration sensors, etc., for collecting information on the vehicle's surrounding environment.

[0065] Computer vision technology: through image processing and analysis, extract useful information, such as vehicle detection, pedestrian detection, etc.

[0066] Sensor fusion technology: fuse the data of multiple sensors to improve the accuracy and reliability of perception.

[0067] Communication technology: realize the communication between vehicles and between vehicles and traffic infrastructure to improve the perception ability of autonomous vehicles.

[0068] For a turning intersection, the zebra crossing is a kind of road semantic information that is relatively easy to perceive and has an auxiliary effect on the generation of the turning reference line of the turning intersection.

[0069] Therefore, as an optional implementation manner, the road semantic information includes: zebra crossing information; step S11 includes: determining a turning start point and a turning end point according to the zebra crossing information and a preset positional relationship, and the preset positional relationship is used to characterize the positional relationships between the turning start point and the turning end point and the zebra crossing respectively.

[0070] In some embodiments, the zebra crossing information includes first zebra crossing information and second zebra crossing information corresponding to the turning direction of a turning intersection. Determining a turning start point and a turning end point according to the zebra crossing information and a preset positional relationship includes: determining the turning start point according to the first zebra crossing information and a preset first positional relationship, where the first positional relationship is used to represent the position of the turning start point relative to the first zebra crossing; determining the turning end point according to the second zebra crossing information and a preset second positional relationship, where the second positional relationship is used to represent the position of the turning end point relative to the second zebra crossing.

[0071] It can be understood that in different traffic scenarios, when the turning direction is different, the required zebra crossing information may be different. Therefore, it is necessary to first obtain the first zebra crossing information and the second zebra crossing information corresponding to the turning direction of the turning intersection. The first zebra crossing information can be understood as the zebra crossing information at the turning start point, and the second zebra crossing information can be understood as the zebra crossing information at the turning end point.

[0072] In some embodiments, the zebra crossing information may include the position of the zebra crossing, the length of the zebra crossing, the width of the zebra crossing, etc.

[0073] In the embodiments of the present disclosure, the positional relationship between the turning start point and the first zebra crossing, and the positional relationship between the turning end point and the second zebra crossing can be preset. In different application scenarios, the positional relationship can be configured in different ways, and it is necessary to ensure that the turning start point is determined based on the first zebra crossing and the turning end point is determined based on the second zebra crossing.

[0074] Please refer to Figure 4 , which is a schematic diagram of the second type of turning information shown according to an exemplary embodiment. As Figure 4 shown, the zebra crossing includes an upper zebra crossing, a left zebra crossing, a right zebra crossing, and a lower zebra crossing. In the case of a right turn, there is a certain distance between the turning start point and the right edge of the lower zebra crossing; there is a certain distance between the turning end point and the lower edge of the right zebra crossing. This certain distance can be configured according to different traffic scenarios, and this certain distance can be: 1 / 4 of the total length of the corresponding zebra crossing, that is, 1 / 4 of the lower zebra crossing; 1 / 4 of the right zebra crossing.

[0075] In this embodiment, as a road semantic information with relatively high accuracy, the zebra crossing can be used to determine the turning start point and the turning end point more precisely.

[0076] In some embodiments, the turning information determined by using the navigation path and the road semantic information respectively can be weighted to integrate more accurate turning information.

[0077] Therefore, the weights corresponding to the navigation path and the weights corresponding to the road semantic information can be preset. After determining the turning information based on the navigation path and the turning information based on the road semantic information, the two types of turning information are weighted and summed. Among them, both the turning start point and the turning end point are represented by position information. Correspondingly, when performing weighted summation, the position data is weighted and summed; for example: the spatial coordinates are weighted and summed.

[0078] In the embodiments of the present disclosure, in addition to the navigation path and the road semantic information, another reference information can also be introduced: the straight path reference line, that is Figure 1 the lane center line shown in, and the straight path reference line can also be obtained by the perception technology introduced above. This kind of reference information can combine the navigation path and / or the road semantic information to determine more accurate turning information.

[0079] In some embodiments, the straight path reference line includes the straight path reference line corresponding to the turning start point and the straight path reference line corresponding to the turning end point. Taking the foregoing Figures 2 - 4 as an example, for the turning start point, the corresponding straight path reference line is the straight path reference line of the current road, and for the turning end point, the corresponding straight path reference line is the straight path reference line of the road after turning. Therefore, it is necessary to use the corresponding straight path reference lines to adjust the turning start point and the turning end point respectively.

[0080] In some embodiments, if the turning information of the turning intersection is determined according to the navigation path of the vehicle or the road semantic information detected by the vehicle, then, for this turning information, the adjustments that can be adopted include: projecting the turning start point onto the corresponding straight path reference line to determine the projected turning start point; projecting the turning end point onto the corresponding straight path reference line to determine the projected turning end point.

[0081] Further, the projected turning start point and the projected turning end point can be directly used as the final turning information; or, according to the projected turning start point and the unprojected turning start point, weighted summation is performed to determine the final turning start point; and, according to the projected turning end point and the unprojected turning end point, weighted summation is performed to determine the final turning end point. Among them, the weight of the projected turning start point is greater than that of the unprojected turning start point, and the weight of the projected turning end point is greater than that of the unprojected turning end point.

[0082] In some embodiments, if the turning information of the turning intersection is determined respectively according to the navigation path of the vehicle and the road semantic information detected by the vehicle, then the straight path reference line can be used to perform separate turning information adjustments, and then final integration is performed.

[0083] Therefore, as an optional implementation, step S11 includes: determining the first turn information based on the navigation path, the first turn information includes the first turn starting point and the first turn end point; determining the second turn information based on the road semantic information, the second turn information includes the second turn starting point and the second turn end point; determining the target turn information based on the first turn information, the second turn information and the straight path reference line, the target turn information includes the target turn starting point and the target turn end point.

[0084] In this implementation, the first turn information is first determined using the navigation path, and the second turn information is determined using the road semantic information; and then the two types of turn information are adjusted using the straight path reference line to determine the target turn information. The first turn information and the second turn information can be determined using the implementation described in the above embodiment, which will not be repeated here.

[0085] In combination with the aforementioned implementation method of the straight path reference line, as an optional implementation method, the target turn information is determined based on the first turn information, the second turn information and the straight path reference line, including: projecting the first turn starting point and the first turn end point onto the corresponding straight path reference line respectively to determine the projected first turn starting point and the projected first turn end point; projecting the second turn starting point and the second turn end point onto the corresponding straight path reference line respectively to determine the projected second turn starting point and the projected second turn end point; determining the target turn starting point based on the first turn starting point, the second turn starting point, the projected first turn starting point and the projected second turn starting point; determining the target turn end point based on the first turn end point, the second turn end point, the projected first turn end point and the projected second turn end point.

[0086] In this embodiment, the first turning starting point and the second turning starting point are both projected onto the straight path reference line corresponding to the turning starting point, and the first turning end point and the second turning end point are both projected onto the straight path reference line corresponding to the turning end point.

[0087] Then, the projected turning starting point and the unprojected turning starting point are combined to determine the target turning starting point; and the projected turning end point and the unprojected turning end point are combined to determine the target turning end point.

[0088] Please refer to Figure 5 , is a schematic diagram of a third type of turning information according to an exemplary embodiment. Figure 5 As shown, Figure 5 Three types of turning starting points are shown in FIG. 1 , the first type is a starting point determined according to a navigation path, the second type is a starting point determined according to a zebra crossing, and the third type is a starting point determined by projection.

[0089] In some embodiments, determining a target turning start point based on a first turning start point, a second turning start point, a projected first turning start point, and a projected second turning start point includes: determining the target turning start point based on the first turning start point, the weight corresponding to the navigation path, the second turning start point, the weight corresponding to the road semantic information, the projected first turning start point, the projected second turning start point, and the weight corresponding to the straight-line path reference line.

[0090] Determining a target turning end point based on a first turning end point, a second turning end point, a projected first turning end point, and a projected second turning end point includes: determining the target turning end point based on the first turning end point, the weight corresponding to the navigation path, the second turning end point, the weight corresponding to the road semantic information, the projected first turning end point, the projected second turning end point, and the weight corresponding to the straight-line path reference line.

[0091] In some embodiments, the weight corresponding to the navigation path is less than the weight corresponding to the road semantic information, and the weight corresponding to the road semantic information may be less than or equal to the weight corresponding to the straight-line path reference line.

[0092] Exemplarily, the target turning start point may be: the first turning start point * the weight corresponding to the navigation path + the second turning start point * the weight corresponding to the road semantic information + the projected second turning start point * the weight corresponding to the straight-line path reference line. The target turning end point may be: the first turning end point * the weight corresponding to the navigation path + the second turning end point * the weight corresponding to the road semantic information + the projected second turning end point * the weight corresponding to the straight-line path reference line.

[0093] Similarly, when performing weighted summation, the position data of the start point or the end point is weighted and summed.

[0094] In some embodiments, for the turning start point, the target turning start point is the result of weighted summation, and it is not necessarily on the straight-line path reference line. For the convenience of path planning, it can be determined on the straight-line path reference line as much as possible.

[0095] Therefore, as an alternative implementation, the reference line generation method further includes: projecting the target turning start point onto the corresponding straight-line path reference line to determine the projected target turning start point; determining the final turning start point based on the projected target turning start point.

[0096] In some embodiments, the projected target turning start point can be directly determined as the final turning start point.

[0097] It can be understood that since the straight-line path reference line corresponding to the turning start point is relatively accurate and matches the straight-line path reference line of the vehicle's current position, a more accurate turning start point can be determined by this re-projection method.

[0098] However, for the turning end point, the corresponding straight path reference line may not be the most accurate reference line. Therefore, in the subsequent process, the turning end point can be gradually optimized.

[0099] In step S12, according to the turning start point and the turning end point included in the turning information, a turning path reference line of the turning intersection can be generated.

[0100] It can be understood that in step S12, a preliminary turning path reference line can be generated first according to the turning start point and the turning end point. After the straight path reference line corresponding to the new turning end point is detected subsequently, the turning end point can be optimized, and then the turning path reference line can be updated according to the optimized turning end point. That is to say, the turning start point does not need to be optimized anymore, and the turning end point can be gradually optimized.

[0101] As an optional implementation manner, step S12 includes: generating a turning path reference line of the turning intersection according to the turning start point, the turning end point, and a preset curve generation algorithm.

[0102] In this implementation manner, since the turning path reference line is a curve, the curve generation algorithm can be used to generate the turning path reference line.

[0103] In some embodiments, the turning information includes the position of the turning start point, the orientation of the turning start point, the position of the turning end point, and the orientation of the turning end point.

[0104] In some embodiments, the orientation of the turning start point and the orientation of the turning end point can be determined by the orientation and position of the navigation path points; for example: select the navigation path points near the turning start point and the turning end point, and determine the respective orientations of the turning start point and the turning end point based on the corresponding navigation path points. In other embodiments, the orientation information can also be determined by other implementation manners.

[0105] In some embodiments, the preset curve generation algorithm can be a second-order Bezier curve algorithm. The second-order Bezier curve is a smooth curve determined by three points, which are the start point, the end point, and the control point respectively. Therefore, the turning start point can be the start point of the second-order Bezier curve, the turning end point can be the end point of the second-order Bezier curve, and the control point can be determined according to the turning start point and the turning end point.

[0106] In some embodiments, the control point of the second-order Bezier curve algorithm is the intersection of the turning start point line and the turning end point line. The turning start point line is determined according to the turning start point and the first orientation, and the turning end point line is determined according to the turning end point and the second orientation.

[0107] In some embodiments, the first orientation may be the orientation of the turning starting point. Thus, along the first orientation, with the turning starting point as the starting point, a turning starting point straight line can be generated. The second orientation may be the orientation of the turning ending point. Thus, along the second orientation, with the turning ending point as the starting point, a turning ending point straight line can be generated. Then, according to the turning starting point straight line and the turning ending point straight line, an intersection point can be generated, and this intersection point is the control point. Furthermore, based on the turning starting point, the turning ending point, and this control point, a second-order Bézier curve can be generated.

[0108] In the embodiments of the present disclosure, a second-order Bézier curve only requires very few control points to generate a relatively complex smooth curve. The starting point position and orientation with good accuracy can just be provided by various features at the intersection turning. Therefore, using a second-order Bézier curve to generate a turning path reference line is not only relatively simple, but also can ensure the accuracy of the result and can be applicable to most scenarios.

[0109] It can be understood that in different traffic scenarios or autonomous driving scenarios, other curve generation algorithms can also be selected according to requirements.

[0110] In some embodiments, the generated turning path reference line can be used to be spliced onto the corresponding straight path reference line (which can be understood as the straight path reference line corresponding to the turning starting point) for the path planning module to plan the turning path. And, in order to maintain the stability of turning at the intersection, after the vehicle's position passes the turning starting point, the spliced reference line can be fixed first. After the vehicle turns to a certain position and the new straight path reference line corresponding to the turning ending point is obtained according to the perception technology, the turning ending point can be adjusted to realize the adjustment of the turning path reference line. So that the adjusted turning path reference line can be connected to the straight path reference line after driving out of the turning intersection.

[0111] Therefore, as an alternative implementation manner, the reference line generation method further includes: in response to the vehicle detecting new road semantic information, determining a new turning ending point according to the new road semantic information; and adjusting the turning path reference line according to the new turning ending point to obtain an adjusted turning path reference line.

[0112] In this implementation manner, the new road semantic information may be the straight path reference line corresponding to the turning ending point, or other information that can be used to adjust the position of the turning ending point.

[0113] In some embodiments, if the new road semantic information is the straight path reference line, then the new turning ending point can be determined according to the projection of the current turning ending point on this straight path reference line, that is, determined according to the projected turning ending point.

[0114] In some embodiments, the new turning ending point can directly be the projected turning ending point.

[0115] In some other embodiments, the turning end point can be adjusted in a progressive manner. At this time, the new turning end point can be determined by combining the projected turning end point and the unprojected turning end point, so that the new turning end point is slowly adjusted to prevent excessive adjustment and cause other risks, such as sudden changes in path planning.

[0116] Therefore, as an optional implementation manner, adjusting the turning path reference line according to the new turning end point to obtain the adjusted turning path reference line includes: determining the adjusted turning end point according to the new turning end point, the weight corresponding to the new turning end point, the turning end point, and the weight corresponding to the turning end point; and determining the adjusted turning path reference line according to the turning start point and the adjusted turning end point.

[0117] In some embodiments, the adjusted turning end point is: the new turning end point * the weight corresponding to the new turning end point + the turning end point * the weight corresponding to the turning end point; where the turning end point can be understood as the turning end point to be adjusted.

[0118] Therefore, the above adjustment method can be expressed as: the end point of the current frame = weight 1 * the end point of the previous frame + weight 2 * the target end point. Where the end point of the current frame is the adjusted turning end point; the end point of the previous frame is the turning end point to be adjusted, and weight 1 is the weight of the end point of the previous frame; the target end point is the new turning end point, that is, the end point determined according to the new road semantic information, and weight 2 is the weight of the target end point.

[0119] Please refer to Figure 6 , which is a schematic diagram of turning end point adjustment shown according to an exemplary embodiment. As Figure 6 shown, the turning end point of the current intersection turning reference line is not projected onto the corresponding straight path reference line. Therefore, when the corresponding straight path reference line is detected, the target end point can be determined, and then the intersection turning reference line can be adjusted based on the target end point.

[0120] Exemplarily, weight 1 can be 0.8, weight 2 can be 0.2, and the sum of weight 1 and weight 2 is 1.

[0121] In some embodiments, the weight corresponding to the new turning end point is less than the weight corresponding to the turning end point; that is, the weight of the end point determined according to the new road semantic information is less than the weight of the turning end point to be adjusted. By setting the weights in this way, a slow and progressive adjustment of the turning end point can be achieved.

[0122] Furthermore, based on the adjusted turning path reference line, the path planning module can update the corresponding turning path. When the vehicle turns successfully, for example, after passing the end point of the turning path reference line and straightening up, the turning path reference line can be removed.

[0123] As can be seen from the introduction of the embodiments of the present disclosure, by using the navigation path of the vehicle and / or the road semantic information detected by the vehicle, the turning information of the turning intersection is determined. By using this turning information, a turning path reference line can be generated, and this turning path reference line can be used to implement the planning of the turning path. Since the generation of this turning path reference line does not rely on a high-precision map, for a vehicle, it is also possible to generate a turning path reference line in the absence of a high-precision map. Thus, this technical solution can be applied to more autonomous driving scenarios or more traffic scenarios, improving the safety and stability of autonomous driving.

[0124] Figure 7 is a block diagram of a reference line generation device shown according to an exemplary embodiment. Referring to Figure 7 , the device includes:

[0125] A determination module 701, configured to determine the turning information of the turning intersection according to the navigation path of the vehicle and / or the road semantic information detected by the vehicle, where the turning information includes a turning start point and a turning end point;

[0126] A generation module 702, configured to generate a turning path reference line of the turning intersection according to the turning start point and the turning end point.

[0127] In an exemplary embodiment, the determination module 701 is further configured to: obtain the path point orientation information of the navigation path, where the path point orientation information includes the orientations corresponding to multiple navigation path points respectively; according to the path point orientation information, determine a first navigation path point and a second navigation path point, where the orientation of the first navigation path point meets a preset turning start point condition, and the orientation of the second navigation path point meets a preset turning end point condition; determine the turning start point according to the first navigation path point, and determine the turning end point according to the second navigation path point.

[0128] In an exemplary embodiment, the determination module 701 is further configured to: determine the turning start point and the turning end point according to the zebra crossing information and a preset positional relationship, where the preset positional relationship is used to characterize the positional relationship between the turning start point and the turning end point and the zebra crossing respectively.

[0129] In an exemplary embodiment, the determination module 701 is further configured to: determine the turning start point according to the first zebra crossing information and a preset first positional relationship, where the first positional relationship is used to characterize the position of the turning start point relative to the first zebra crossing; determine the turning end point according to the second zebra crossing information and a preset second positional relationship, where the second positional relationship is used to characterize the position of the turning end point relative to the second zebra crossing.

[0130] In an exemplary embodiment, the determining module 701 is further configured to: determine first turning information according to the navigation path, where the first turning information includes a first turning start point and a first turning end point; determine second turning information according to the road semantic information, where the second turning information includes a second turning start point and a second turning end point; determine target turning information according to the first turning information, the second turning information, and the straight-line path reference line, where the target turning information includes a target turning start point and a target turning end point.

[0131] In an exemplary embodiment, the determining module 701 is further configured to: project the first turning start point and the first turning end point onto the corresponding straight-line path reference line respectively to determine a projected first turning start point and a projected first turning end point; project the second turning start point and the second turning end point onto the corresponding straight-line path reference line respectively to determine a projected second turning start point and a projected second turning end point; determine the target turning start point according to the first turning start point, the second turning start point, the projected first turning start point, and the projected second turning start point; determine the target turning end point according to the first turning end point, the second turning end point, the projected first turning end point, and the projected second turning end point.

[0132] In an exemplary embodiment, the determining module 701 is further configured to: determine the target turning start point according to the first turning start point, the weight corresponding to the navigation path, the second turning start point, the weight corresponding to the road semantic information, the projected first turning start point, the projected second turning start point, and the weight corresponding to the straight-line path reference line; determine the target turning end point according to the first turning end point, the weight corresponding to the navigation path, the second turning end point, the weight corresponding to the road semantic information, the projected first turning end point, the projected second turning end point, and the weight corresponding to the straight-line path reference line.

[0133] In an exemplary embodiment, the determining module 701 is further configured to: project the target turning start point onto the corresponding straight-line path reference line to determine a projected target turning start point; determine the final turning start point according to the projected target turning start point.

[0134] In an exemplary embodiment, the generating module 702 is further configured to: generate a turning path reference line of the turning intersection according to the turning start point, the turning end point, and a preset curve generation algorithm.

[0135] In an exemplary embodiment, the preset curve generation algorithm is a second-order Bezier curve algorithm. The control points of the second-order Bezier curve algorithm are the intersection points of the straight line at the starting point of the turn and the straight line at the ending point of the turn. The straight line at the starting point of the turn is determined according to the starting point of the turn and the first orientation, and the straight line at the ending point of the turn is determined according to the ending point of the turn and the second orientation.

[0136] In an exemplary embodiment, the generation module 702 is further configured to: in response to the vehicle detecting new road semantic information, determine a new ending point of the turn according to the new road semantic information; and adjust the turn path reference line according to the new ending point of the turn to obtain an adjusted turn path reference line.

[0137] In an exemplary embodiment, the generation module 702 is further configured to: determine an adjusted ending point of the turn according to the new ending point of the turn, the weight corresponding to the new ending point of the turn, the ending point of the turn, and the weight corresponding to the ending point of the turn; and determine an adjusted turn path reference line according to the starting point of the turn and the adjusted ending point of the turn.

[0138] In an exemplary embodiment, the weight corresponding to the new ending point of the turn is less than the weight corresponding to the ending point of the turn.

[0139] Regarding the device in the above embodiments, the specific manners in which each module performs operations have been described in detail in the embodiments related to the method, and will not be elaborated herein.

[0140] The present disclosure also provides a computer-readable storage medium, on which computer program instructions are stored. When the program instructions are executed by a processor, the steps of the reference line generation method provided by the present disclosure are implemented.

[0141] Figure 8 FIG. is a block diagram of a vehicle 800 shown according to an exemplary embodiment. For example, the vehicle 800 may be a hybrid vehicle, or may also be a non-hybrid vehicle, an electric vehicle, a fuel cell vehicle, or other types of vehicles. The vehicle 800 may be an autonomous vehicle, a semi-autonomous vehicle, or a non-autonomous vehicle.

[0142] Referring to Figure 8 , the vehicle 800 may include various subsystems. For example, an infotainment system 810, a perception system 820, a decision control system 830, a drive system 840, and a computing platform 850. Among them, the vehicle 800 may further include more or fewer subsystems, and each subsystem may include multiple components. In addition, each subsystem and each component of the vehicle 800 may be interconnected in a wired or wireless manner.

[0143] In some embodiments, the infotainment system 810 may include a communication system, an entertainment system, a navigation system, and the like.

[0144] The perception system 820 may include several sensors for sensing information about the environment around the vehicle 800. For example, the perception system 820 may include a global positioning system (the global positioning system may be a GPS system, a Beidou system, or other positioning systems), an inertial measurement unit (IMU), lidar, millimeter-wave radar, ultrasonic radar, and a camera device.

[0145] The decision control system 830 may include a computing system, a vehicle controller, a steering system, an accelerator, and a braking system.

[0146] The drive system 840 may include components that provide motive power for the vehicle 800. In one embodiment, the drive system 840 may include an engine, an energy source, a powertrain, and wheels. The engine may be one or a combination of an internal combustion engine, an electric motor, and an air compression engine. The engine can convert the energy provided by the energy source into mechanical energy.

[0147] Some or all of the functions of the vehicle 800 are controlled by the computing platform 850. The computing platform 850 may include at least one processor 851 and a memory 852, and the processor 851 may execute instructions 853 stored in the memory 852.

[0148] The processor 851 may be any conventional processor, such as a commercially available CPU. The processor may also include, for example, a Graphic Process Unit (GPU), a Field Programmable Gate Array (FPGA), a System on Chip (SOC), an Application Specific Integrated Circuit (ASIC), or a combination thereof.

[0149] The memory 852 may be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, a magnetic disk, or an optical disk.

[0150] In addition to instruction 853, the memory 852 may also store data, such as road maps, route information, data on the position, direction, speed, etc. of the vehicle. The data stored in the memory 852 can be used by the computing platform 850.

[0151] In an embodiment of the present disclosure, the processor 851 may execute the instruction 853 to complete all or part of the steps of the above-mentioned reference line generation method.

[0152] In another exemplary embodiment, a computer program product is also provided. The computer program product includes a computer program that can be executed by a programmable device. The computer program has a code portion for executing the above-mentioned reference line generation method when executed by the programmable device.

[0153] Those skilled in the art will readily conceive of other embodiments of the present disclosure after considering the specification and practicing the present disclosure. This application is intended to cover any variations, uses, or adaptations of the present disclosure, which follow the general principles of the present disclosure and include known common knowledge or conventional technical means in the technical field not disclosed in the present disclosure. The specification and embodiments are only regarded as exemplary, and the true scope and spirit of the present disclosure are pointed out by the following claims.

[0154] It should be understood that the present disclosure is not limited to the exact structures described above and shown in the drawings, and various modifications and changes can be made without departing from its scope. The scope of the present disclosure is only limited by the appended claims.

Claims

1. A reference line generation method, characterized in that, Including: Determine the turning information of a turning intersection according to the navigation path of the vehicle and / or the road semantic information detected by the vehicle, where the turning information includes a turning start point and a turning end point; Generate a turning path reference line of the turning intersection according to the turning start point and the turning end point.

2. The reference line generation method according to claim 1, wherein The determining the turning information of the turning intersection according to the navigation path of the vehicle and / or the road semantic information detected by the vehicle includes: Obtain the path point orientation information of the navigation path, where the path point orientation information includes the orientations respectively corresponding to multiple navigation path points; Determine a first navigation path point and a second navigation path point according to the path point orientation information, where the orientation of the first navigation path point meets a preset turning start point condition, and the orientation of the second navigation path point meets a preset turning end point condition; Determine the turning start point according to the first navigation path point, and determine the turning end point according to the second navigation path point.

3. The reference line generation method according to claim 1, wherein The road semantic information includes: zebra crossing information. The determining the turning information of the turning intersection according to the navigation path of the vehicle and / or the road semantic information detected by the vehicle includes: Determine the turning start point and the turning end point according to the zebra crossing information and a preset positional relationship, where the preset positional relationship is used to represent the positional relationships between the turning start point and the turning end point and the zebra crossing respectively.

4. The reference line generation method according to claim 3, characterized in that The zebra crossing information includes first zebra crossing information and second zebra crossing information corresponding to the turning direction of the turning intersection. The determining the turning start point and the turning end point according to the zebra crossing information and a preset positional relationship includes: Determine the turning start point according to the first zebra crossing information and a preset first positional relationship, where the first positional relationship is used to represent the position of the turning start point relative to the first zebra crossing; Determine the turning end point according to the second zebra crossing information and a preset second positional relationship, where the second positional relationship is used to represent the position of the turning end point relative to the second zebra crossing.

5. The reference line generation method according to claim 1, wherein The determining the turning information of the turning intersection according to the navigation path of the vehicle and / or the road semantic information detected by the vehicle includes: Determine first turning information according to the navigation path, where the first turning information includes a first turning start point and a first turning end point; Determine second turning information according to the road semantic information, where the second turning information includes a second turning start point and a second turning end point; Determine target turning information according to the first turning information, the second turning information and a straight path reference line, where the target turning information includes a target turning start point and a target turning end point.

6. The reference line generation method according to claim 5, characterized in that The straight path reference line includes a straight path reference line corresponding to the turning start point and a straight path reference line corresponding to the turning end point. The determining the target turning information according to the first turning information, the second turning information and the straight path reference line includes: Project the first turning start point and the first turning end point onto the corresponding straight path reference lines respectively to determine the projected first turning start point and the projected first turning end point; Project the second turning start point and the second turning end point onto the corresponding straight path reference lines respectively to determine the projected second turning start point and the projected second turning end point; Determine the target turning start point according to the first turning start point, the second turning start point, the projected first turning start point and the projected second turning start point; Determine the target turning end point according to the first turning end point, the second turning end point, the projected first turning end point and the projected second turning end point.

7. The reference line generation method according to claim 6, characterized in that, The step of determining the target turning start point according to the first turning start point, the second turning start point, the projected first turning start point and the projected second turning start point includes: Determine the target turning start point according to the first turning start point, the weight corresponding to the navigation path, the second turning start point, the weight corresponding to the road semantic information, the projected first turning start point, the projected second turning start point and the weight corresponding to the straight path reference line; The step of determining the target turning end point according to the first turning end point, the second turning end point, the projected first turning end point and the projected second turning end point includes: Determine the target turning end point according to the first turning end point, the weight corresponding to the navigation path, the second turning end point, the weight corresponding to the road semantic information, the projected first turning end point, the projected second turning end point and the weight corresponding to the straight path reference line.

8. The reference line generation method according to claim 7, characterized in that, The reference line generation method further includes: Project the target turning start point onto the corresponding straight path reference line to determine the projected target turning start point; Determine the final turning start point according to the projected target turning start point.

9. The reference line generation method according to claim 1, wherein The step of generating the turning path reference line of the turning intersection according to the turning start point and the turning end point includes: Generate the turning path reference line of the turning intersection according to the turning start point, the turning end point and a preset curve generation algorithm.

10. The reference line generation method according to claim 9, characterized in that The preset curve generation algorithm is a second-order Bezier curve algorithm, and the control point of the second-order Bezier curve algorithm is the intersection of the turning start point straight line and the turning end point straight line. The turning start point straight line is determined according to the turning start point and the first orientation, and the turning end point straight line is determined according to the turning end point and the second orientation.

11. The reference line generation method according to claim 1, wherein The reference line generation method further includes: In response to the vehicle detecting new road semantic information, determine a new turning end point according to the new road semantic information; Adjust the turning path reference line according to the new turning end point to obtain an adjusted turning path reference line.

12. The reference line generation method according to claim 11, wherein The step of adjusting the turning path reference line according to the new turning end point to obtain an adjusted turning path reference line includes: Determine the adjusted turning end point according to the new turning end point, the weight corresponding to the new turning end point, the turning end point and the weight corresponding to the turning end point; Determine the adjusted turning path reference line according to the turning start point and the adjusted turning end point.

13. The reference line generation method according to claim 12, characterized in that, The weight corresponding to the new turning end point is less than the weight corresponding to the turning end point.

14. A reference line generation device, characterized in that, including: A determination module, configured to determine turning information of a turning intersection according to a navigation path of a vehicle and / or road semantic information detected by the vehicle, where the turning information includes a turning start point and a turning end point; A generation module, configured to generate a turning path reference line of the turning intersection according to the turning start point and the turning end point.

15. A vehicle, characterized in that, Comprising: A processor; A memory for storing processor-executable instructions; Wherein, the processor is configured to execute the steps of the method according to any one of claims 1-13.

16. A computer-readable storage medium having computer program instructions stored thereon, characterized in that, When the program instruction is executed by the processor, the steps of the method according to any one of claims 1-13 are implemented.

Images