Method, device and electronic equipment for generating a predicted driving trajectory of an autonomous vehicle

Abstract

This disclosure provides a method, apparatus, and electronic device for generating predicted driving trajectories for autonomous vehicles, relating to the fields of artificial intelligence, particularly autonomous driving, computer science, and map data processing. The specific implementation involves: acquiring at least one predicted driving trajectory for the autonomous vehicle; determining at least one first predicted driving trajectory that conforms to target driving rules from the at least one predicted driving trajectory, wherein the target driving rules guide the autonomous vehicle to drive normally on a road segment; determining a target lane corresponding to the first predicted driving trajectory on the road segment based on trajectory points on the first predicted driving trajectory, wherein the target lane is the lane the autonomous vehicle is to be driven to and conforms to the target driving rules; and generating a second predicted driving trajectory for the autonomous vehicle based on the trajectory points and the target lane, wherein the second predicted driving trajectory controls the autonomous vehicle to drive to the target lane.

Description

Technical Field

[0001] This disclosure relates to the field of artificial intelligence technology, and in particular to a method, apparatus and electronic device for generating a predicted driving trajectory of an autonomous vehicle in the fields of autonomous driving, computer science and map data processing. Background Technology

[0002] Currently, autonomous vehicles on the market generate trajectory points directly through trajectory prediction models to predict their driving trajectory. However, the driving trajectory predicted by the above method often shows obvious mismatches, making it impossible to accurately control the autonomous vehicle and thus causing safety hazards during the driving process. Summary of the Invention

[0003] This disclosure provides a method, apparatus, and electronic device for generating predicted driving trajectories for autonomous vehicles.

[0004] According to another aspect of this disclosure, a method for generating a predicted driving trajectory for an autonomous vehicle is provided. The method may include: acquiring at least one predicted driving trajectory of the autonomous vehicle; determining at least one first predicted driving trajectory that conforms to a target driving rule among the at least one predicted driving trajectory, wherein the target driving rule is used to guide the autonomous vehicle to drive in a normal driving state on a road segment; determining a target lane corresponding to the first predicted driving trajectory on the road segment based on trajectory points on the first predicted driving trajectory, wherein the target lane is a lane that the autonomous vehicle is to drive to and that conforms to the target driving rule; and generating a second predicted driving trajectory of the autonomous vehicle based on the trajectory points and the target lane, wherein the second predicted driving trajectory is used to control the autonomous vehicle to drive to the target lane.

[0005] According to another aspect of this disclosure, another method for generating a predicted driving trajectory for an autonomous vehicle is also provided. This method may include: displaying at least one predicted driving trajectory output by a trajectory prediction model on an operating interface, wherein the trajectory prediction model is used to predict the driving trajectory of the autonomous vehicle; displaying at least one first predicted driving trajectory from the at least one predicted driving trajectory that conforms to a target driving rule on the operating interface, wherein the target driving rule is used to guide the autonomous vehicle to drive in a normal driving state on a road segment; and, in response to a trajectory generation command applied to the operating interface, displaying a second predicted driving trajectory from the at least one first predicted driving trajectory on the operating interface, wherein the second predicted driving trajectory is used to control the autonomous vehicle to drive to a target lane, and is generated based on trajectory points on the first predicted driving trajectory and the target lane, wherein the target lane is a lane to which the autonomous vehicle is to drive and conforms to the target driving rule, and the target lane is determined on the road segment based on trajectory points on the first predicted driving trajectory.

[0006] According to another aspect of this disclosure, a predictive driving trajectory generation apparatus for an autonomous vehicle is also provided. The apparatus may include: an acquisition unit for acquiring at least one predicted driving trajectory of the autonomous vehicle; a first determination unit for determining at least one first predicted driving trajectory conforming to a target driving rule from among the at least one predicted driving trajectory, wherein the target driving rule is used to guide the autonomous vehicle to drive in a normal driving state on a road segment; a second determination unit for determining a target lane corresponding to the first predicted driving trajectory on the road segment based on trajectory points on the first predicted driving trajectory, wherein the target lane is a lane to which the autonomous vehicle is to be driven and conforms to the target driving rule; and a generation unit for generating a second predicted driving trajectory of the autonomous vehicle based on the trajectory points and the target lane, wherein the second predicted driving trajectory is used to control the autonomous vehicle to drive to the target lane.

[0007] According to another aspect of this disclosure, another predictive driving trajectory generation device for an autonomous vehicle is also provided. The device may include: a first display unit for displaying on an operating interface at least one predicted driving trajectory output by a trajectory prediction model, wherein the trajectory prediction model is used to predict the driving trajectory of the autonomous vehicle; a second display unit for displaying on the operating interface at least one first predicted driving trajectory that conforms to a target driving rule, wherein the target driving rule is used to guide the autonomous vehicle to drive in a normal driving state on a road segment; and a third display unit for responding to a trajectory generation command applied to the operating interface and displaying on the operating interface a second predicted driving trajectory among the at least one first predicted driving trajectory, wherein the second predicted driving trajectory is used to control the autonomous vehicle to drive to a target lane, and is generated based on trajectory points on the first predicted driving trajectory and the target lane, wherein the target lane is the lane to which the autonomous vehicle is to drive and conforms to the target driving rule, and the target lane is determined on the road segment based on the trajectory points on the first predicted driving trajectory.

[0008] According to one aspect of this disclosure, an electronic device is also provided. The electronic device may include: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the predicted driving trajectory generation method for an autonomous vehicle according to embodiments of this disclosure.

[0009] According to another aspect of this disclosure, a non-transitory computer-readable storage medium storing computer instructions is also provided, wherein the computer instructions are used to cause a computer to execute the method for generating a predicted driving trajectory of an autonomous vehicle according to embodiments of this disclosure.

[0010] According to another aspect of this disclosure, a computer program product is also provided, which may include a computer program that, when executed by a processor, implements the method for generating a predicted driving trajectory of an autonomous vehicle according to embodiments of this disclosure.

[0011] According to another aspect of this disclosure, an autonomous vehicle is also provided, which may include the aforementioned electronic equipment.

[0012] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0013] The accompanying drawings are for better illustration of this scheme and do not constitute a limitation of this disclosure. Wherein:

[0014] Figure 1 This is a flowchart of a method for generating a predicted driving trajectory of an autonomous vehicle according to an embodiment of the present disclosure;

[0015] Figure 2a This is a flowchart of another method for generating a predicted driving trajectory for an autonomous vehicle according to an embodiment of the present disclosure;

[0016] Figure 2b This is a schematic diagram of an operation interface according to an embodiment of the present disclosure;

[0017] Figure 3 This is a flowchart of a driving trajectory generation method according to an embodiment of the present disclosure;

[0018] Figure 4 This is a schematic diagram of a trajectory point outside an intersection according to an embodiment of the present disclosure;

[0019] Figure 5 This is a schematic diagram of a trajectory point crossing a boundary according to an embodiment of the present disclosure;

[0020] Figure 6 This is a schematic diagram of trajectory point retraction according to an embodiment of the present disclosure;

[0021] Figure 7 This is a schematic diagram of a target lane selection according to an embodiment of the present disclosure;

[0022] Figure 8 This is a schematic diagram of the orientation of an autonomous vehicle according to an embodiment of the present disclosure;

[0023] Figure 9 This is a schematic diagram of a target lane selection according to an embodiment of the present disclosure;

[0024] Figure 10This is a schematic diagram illustrating the selection of constraint points according to an embodiment of the present disclosure;

[0025] Figure 11a This is a schematic diagram of constraint point backtracking according to an embodiment of the present disclosure;

[0026] Figure 11b This is a schematic diagram of the generation result of a trajectory line according to an embodiment of the present disclosure;

[0027] Figure 11c This is a schematic diagram of the post-processing result of a trajectory line according to an embodiment of the present disclosure;

[0028] Figure 12 This is a schematic diagram of a predictive driving trajectory generation device for an autonomous vehicle according to an embodiment of the present disclosure;

[0029] Figure 13 This is a schematic diagram of another predictive driving trajectory generation device for an autonomous vehicle according to an embodiment of the present disclosure;

[0030] Figure 14 This is a block diagram of an electronic device for generating a predicted driving trajectory of an autonomous vehicle according to an embodiment of the present disclosure. Detailed Implementation

[0031] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments to aid understanding, and should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0032] This disclosure provides a method for generating a predicted driving trajectory for an autonomous vehicle. Figure 1 This is a flowchart of a method for generating a predicted driving trajectory for an autonomous vehicle according to an embodiment of the present disclosure, as shown below. Figure 1 As shown, an implementation scheme for the method of generating the predicted driving trajectory of the autonomous vehicle may include at least the following implementation steps:

[0033] Step S102: Obtain at least one predicted driving trajectory of the autonomous vehicle.

[0034] In the technical solution provided by step S102 disclosed in this application, at least one predicted driving trajectory of an autonomous vehicle can be obtained. The predicted driving trajectory can be the predicted trajectory line of the autonomous vehicle or the driving path of the autonomous vehicle.

[0035] For example, this embodiment can use a trajectory prediction model to predict the driving trajectory of an autonomous vehicle, obtaining at least one predicted driving trajectory. The trajectory prediction model can be used to predict the trajectory of an autonomous vehicle. It should be noted that the at least one predicted driving trajectory can be obtained by the trajectory prediction model; this is merely an example and does not impose specific limitations on the method of obtaining the predicted driving trajectory.

[0036] Step S104: Among at least one predicted driving trajectory, determine at least one first predicted driving trajectory that conforms to the target driving rule, wherein the target driving rule is used to guide the autonomous vehicle to drive in a normal driving state on the road segment.

[0037] In the technical solution provided by step S104 of this application, at least one first predicted driving trajectory that conforms to the target driving rule can be determined from at least one predicted driving trajectory. The target driving rule can be prior knowledge and / or map information, which can be used to filter at least one predicted driving trajectory and guide the autonomous vehicle to drive in a normal driving state on the road segment. Prior knowledge can be expert prior knowledge, which can be some rules or experiences that are relatively easy to judge, such as a trajectory that requires a left turn on a dedicated left-turn lane, a trajectory that likely requires slowing down when encountering a red light, or a trajectory that cannot be reversed due to real obstacles. This is only an example and does not impose specific limitations on the content of prior knowledge. Map information can be information from various road segments acquired in advance, such as information on road boundaries and center lines in various road segments. This is only an example and does not impose specific limitations on the acquisition method and content of map information. Normal driving state can refer to a driving state without violations of traffic regulations such as illegal driving or driving against traffic. This is only an example and does not impose specific limitations on the normal driving state. The first predicted driving trajectory can be a trajectory line that conforms to the autonomous driving rule from at least one predicted driving trajectory.

[0038] Optionally, to obtain at least one predicted driving trajectory, the at least one predicted driving trajectory can be filtered based on the target driving rules to determine the predicted driving trajectory of the normal driving state of the autonomous vehicle from the at least one predicted driving trajectory, thereby obtaining at least one first predicted driving trajectory that conforms to the target driving rules.

[0039] For example, at least one predicted driving trajectory is obtained on road A, namely predicted driving trajectory B, predicted driving trajectory C, predicted driving trajectory D, predicted driving trajectory E, and predicted driving trajectory F. The obtained at least one predicted driving trajectory can be judged based on prior knowledge and map information. It is determined that predicted driving trajectory B has a problem where trajectory points are outside the intersection, but vehicles cannot drive outside the road during operation. Therefore, predicted driving trajectory B is an abnormal driving state and does not conform to the target driving rules.

[0040] Optionally, if multiple trajectory points in predicted driving trajectory C cross the yellow road line, it constitutes a traffic violation. Therefore, predicted driving trajectory C is an abnormal driving state and does not conform to the target driving rules. Similarly, predicted driving trajectory D exhibits trajectory point reversal, which also constitutes a traffic violation. Therefore, predicted driving trajectory D is also an abnormal driving state and does not conform to the target driving rules. Optionally, if predicted driving trajectories E and F conform to the target driving rules, then at least one first predicted driving trajectory can be determined as predicted driving trajectory E and predicted driving trajectory F.

[0041] In this embodiment of the disclosure, at least one predicted driving trajectory of an autonomous vehicle is obtained, and at least one first predicted driving trajectory in the normal driving state among the at least one predicted driving trajectory is determined by the target driving rule. That is, the predicted driving trajectory is filtered by the target driving rule to further determine a reasonable (satisfying the target driving rule) driving trajectory to ensure the normal driving of the autonomous vehicle.

[0042] Step S106: Based on the trajectory points on the first predicted driving trajectory, determine the target lane corresponding to the first predicted driving trajectory on the road segment, wherein the target lane is the lane that the autonomous vehicle is to drive to and that conforms to the target driving rules.

[0043] In the technical solution provided by step S106 of this application, a target lane corresponding to the first predicted driving trajectory can be determined on a road segment based on the trajectory points on the first predicted driving trajectory. The trajectory points on the first predicted driving trajectory can be reasonable trajectory points. The target lane can be the lane to which the autonomous vehicle is to travel, for example, it can be the exit lane of the autonomous vehicle, and it can be a lane that conforms to the target driving rules.

[0044] For example, the orientation of the last trajectory point can be determined based on the trajectory points on the first predicted driving trajectory. Based on the orientation of the last trajectory point, the target lane corresponding to the first predicted driving trajectory and conforming to the target driving rules can be determined on the road segment, and the autonomous vehicle can drive in the target lane.

[0045] It should be noted that the above method for determining the target lane is only an example, and any method for determining the target lane based on trajectory points should be within the protection scope of this disclosure.

[0046] Step S108: Based on the trajectory points and the target lane, a second predicted driving trajectory for the autonomous vehicle is generated, wherein the second predicted driving trajectory is used to control the autonomous vehicle to drive to the target lane.

[0047] In the technical solution provided by step S108 of this application, a second predicted driving trajectory of the autonomous vehicle can be generated based on trajectory points and the target lane. The second predicted driving trajectory can be a driving trajectory that conforms to the target driving rules and can be used to control the autonomous vehicle to drive onto the target lane.

[0048] Optionally, this embodiment can determine a first predicted driving trajectory that conforms to the target driving rules, determine the target lane of the autonomous vehicle based on the trajectory points of the first predicted driving rules, and generate a second predicted driving trajectory of the autonomous vehicle based on the target lane and trajectory points. The autonomous vehicle can be precisely controlled to drive to the target lane through the second predicted driving trajectory.

[0049] Since this embodiment of the disclosure takes into account the different driving requirements and conditions on different road sections, a reasonable first predicted driving trajectory is determined by using target driving rules, thereby improving the rationality of the predicted driving trajectory. Furthermore, a target lane is determined based on the trajectory points of the first predicted driving trajectory, and a second predicted driving trajectory is determined based on the trajectory points and the target lane, thereby improving the adaptability of the predicted driving trajectory to the target lane and further ensuring vehicle driving safety.

[0050] Through steps S102 to S108, at least one predicted driving trajectory of the autonomous vehicle is obtained. Among the at least one predicted driving trajectory, at least one first predicted driving trajectory conforming to the target driving rules is determined, wherein the target driving rules guide the autonomous vehicle to drive normally on the road segment. Based on the trajectory points on the first predicted driving trajectory, a target lane corresponding to the first predicted driving trajectory is determined on the road segment, wherein the target lane is the lane the autonomous vehicle is to be driven to and conforms to the target driving rules. Based on the trajectory points and the target lane, a second predicted driving trajectory of the autonomous vehicle is generated, wherein the second predicted driving trajectory is used to control the autonomous vehicle to drive to the target lane. In other words, this embodiment of the present disclosure determines a reasonable first predicted driving trajectory conforming to the target driving rules from at least one predicted driving trajectory of the autonomous vehicle, determines the corresponding target lane (exit lane) based on the reasonable trajectory points on the first predicted driving trajectory, and generates a second predicted driving trajectory conforming to the target driving rules based on the reasonable trajectory points and the target lane, thereby achieving the technical effect of effectively generating the driving trajectory of the autonomous vehicle and solving the technical problem of not being able to effectively generate the driving trajectory of the autonomous vehicle.

[0051] The method described in this embodiment will now be described in further detail.

[0052] As an optional embodiment, step S104, the target driving rule is determined by the map information of the road segment, and determining at least one first predicted driving trajectory that conforms to the target driving rule in at least one predicted driving trajectory includes: in at least one predicted driving trajectory, determining the predicted driving trajectory that matches the map information of the road segment as the first predicted driving trajectory.

[0053] In this embodiment, the target driving rule can be determined by the map information of the road segment. Among at least one predicted driving trajectory, the predicted trajectory that matches the map information of the road segment can be determined as the first predicted driving trajectory. The map information can be information from each road segment that has been acquired in advance, such as information on road boundaries and median lines in each road segment. This is only an example and no specific restrictions are placed on the method and content of map information acquisition.

[0054] Optionally, based on the map information of the road segment, at least one predicted driving trajectory can be filtered to determine the predicted driving trajectory that matches the map information of the road segment among the at least one predicted driving trajectory, and the above-mentioned predicted driving trajectory can be determined as the first predicted driving trajectory.

[0055] For example, map information of each road segment can be obtained in advance and stored. At least one predicted driving trajectory can be obtained, map information of the road segments of the at least one predicted driving trajectory can be determined, and a target driving rule can be obtained. The at least one predicted driving trajectory can be processed based on the target driving rule to determine the predicted driving trajectory that matches the target driving rule among the at least one predicted driving trajectory, so as to obtain the first predicted driving trajectory.

[0056] In this embodiment of the disclosure, by using map information of road segments to match at least one predicted driving trajectory, the technical problem in the related art of low drivability of the predicted driving trajectory due to mismatch between the predicted driving trajectory and map information is solved, thereby achieving the technical effect of improving the drivability of the predicted driving trajectory.

[0057] As an optional embodiment, step S104, the target driving rule is determined by the prior driving information of the autonomous vehicle, and determining at least one first predicted driving trajectory that conforms to the target driving rule in at least one predicted driving trajectory includes: in at least one predicted driving trajectory, determining the predicted driving trajectory that matches the prior driving information as the first predicted driving trajectory.

[0058] In this embodiment, the target driving rule can be determined by the prior driving information of the autonomous vehicle. From at least one predicted driving trajectory, a predicted driving trajectory that matches the prior driving information can be determined, and this predicted driving trajectory can be identified as the first predicted driving trajectory. The prior driving information can refer to rules or experiences that are relatively easy for humans to judge, such as requiring a left turn in a dedicated left-turn lane, the high probability of needing to slow down when encountering a red light, or a trajectory where a real obstacle prevents the user from retreating. This is merely an example and does not impose specific limitations on the content of the prior driving information. Prior driving information can also be referred to as prior knowledge (or prior information).

[0059] Optionally, prior driving information can be obtained, which may be information about previously acquired driving experience. Based on the prior driving information, at least one predicted driving trajectory can be filtered to determine the predicted driving trajectory that matches the prior driving information among the at least one predicted driving trajectory, and the aforementioned predicted driving trajectory can be determined as the first predicted driving trajectory.

[0060] For example, prior driving information can be obtained and stored in advance. At least one predicted driving trajectory can be obtained, a target driving rule can be determined based on the prior driving information, and the at least one predicted driving trajectory can be processed based on the target driving rule to determine the predicted driving trajectory that matches the target driving rule among the at least one predicted driving trajectory, so as to obtain the first predicted driving trajectory.

[0061] In related technologies, an autonomous vehicle simply selects one of at least one predicted driving trajectory and drives according to that trajectory. However, this method suffers from low matching between the predicted trajectory and the road segment, resulting in ineffective vehicle control. In this embodiment, a target driving rule is determined based on map information and prior driving information of the road segment. The at least one predicted driving trajectory is then filtered based on this target driving rule to obtain a first predicted driving trajectory. Finally, the vehicle's final driving trajectory is determined based on this first predicted driving trajectory. This improves the matching between the predicted driving trajectory and the road segment, thereby enhancing vehicle safety during operation.

[0062] As an optional embodiment, step S106, determining the target lane corresponding to the first predicted driving trajectory based on the trajectory points on the first predicted driving trajectory, includes: determining the target lane based at least on the last trajectory point on the first predicted driving trajectory, wherein the multiple trajectory points on the first predicted driving trajectory are arranged according to the corresponding trajectory directions.

[0063] In this embodiment, the first predicted driving trajectory can be composed of multiple trajectory points, and can be a trajectory obtained by arranging multiple trajectory points according to their corresponding trajectory directions. The target lane can be determined based at least on the last trajectory point on the first predicted driving trajectory.

[0064] Optionally, trajectory points on the first predicted driving trajectory can be determined, and based on the trajectory points, the last trajectory point on the first predicted driving trajectory can be determined, and the target lane can be determined based on the last trajectory point.

[0065] As an optional embodiment, determining the target lane based at least on the end trajectory point on the first predicted driving trajectory includes: determining the candidate target lane as the target lane in response to the last trajectory point being on a candidate target lane; or, in response to the last trajectory point not being on a candidate target lane, determining the target lane based on a first orientation determined by the last two trajectory points of the first predicted driving trajectory and a second orientation of the last trajectory point pointing to each of at least one candidate target lane.

[0066] In this embodiment, it is determined whether the last trajectory point is on a candidate target lane. If it is determined that the last trajectory point is on a candidate target lane, the candidate target lane can be determined as the target lane in response to this. If it is determined that the last trajectory point is not on a candidate target lane, the first orientation of the line segment formed by the last two trajectory points of the first predicted driving trajectory, and the second orientation of the line segment pointing from the last trajectory point to each of the candidate target lanes can be determined in response to this. The target lane can be determined based on the first and second orientations. The candidate target lane can be the actual target lane at the exit, and there can be multiple candidate target lanes. The first orientation (Obs_end_heading) can be the orientation of the line connecting the last two trajectory points of the first predicted driving trajectory. The second orientation (pos_heading) can be the orientation of the line connecting the last trajectory point to the exit point of each candidate target lane.

[0067] In this embodiment, the target lane is determined based on the position of the last trajectory point on the first predicted driving trajectory. When the position of the last trajectory point is on a candidate target lane, the candidate target lane can be determined as the target lane; when the position of the last trajectory point is not on a candidate target lane, the target lane can be determined from the candidate target lanes based on the second orientation of the line connecting the last trajectory point to the exit point of the candidate target lane and the first orientation of the line connecting the last two trajectory points. This improves the efficiency of path planning and solves the problem of low path planning efficiency due to the planned path being unusable.

[0068] Optionally, when determining the target lane, if the last trajectory point has already landed on a candidate target lane, then the candidate target lane can be determined as an exit lane, and the candidate target lane can be designated as the target lane. If the last trajectory point is not in a candidate target lane, then a first orientation can be determined based on the last two trajectory points of the first predicted driving trajectory, a second orientation can be determined based on the last trajectory point, and the target lane can be determined based on the first and second orientations.

[0069] For example, it can be determined whether the last trajectory point of the predicted driving trajectory is on a candidate target lane (which could be the actual exit lane). If the last trajectory point is determined to be on a candidate target lane, the candidate target lane can be designated as the target lane. If the last trajectory point is not on an exit lane, a first orientation can be determined based on the last two trajectory points of the predicted driving trajectory, and a second orientation can be determined based on the last trajectory point pointing to the candidate target lane. The target lane can then be determined based on the first and second orientations.

[0070] As an optional embodiment, determining a target lane based on a first orientation determined by the last two trajectory points of a first predicted driving trajectory, and a second orientation pointing from the last trajectory point to each of at least one candidate target lanes, includes: determining, among at least one candidate target lane, a second orientation with the smallest angle to the first orientation; and determining the candidate target lane corresponding to the determined second orientation as the target lane.

[0071] In this embodiment, if the last trajectory point is not in a candidate target lane, the first orientation of the line segment connecting the last two trajectory points of the first predicted driving trajectory can be determined, along with the second orientation pointing from the last trajectory point to each of the at least one candidate target lanes, thus obtaining at least one second orientation. The angle between the first orientation and each second orientation is determined, and the second orientation with the smallest angle to the first orientation is selected. The candidate target lane corresponding to the determined second orientation is then designated as the target lane. The angle between the second orientation and the first orientation can be used to represent the difference between the first and second orientations.

[0072] Optionally, if the last trajectory point is not in the candidate target lane, a first orientation and a second orientation can be determined. The second orientation, which has the smallest angle with the first orientation, can be determined. The angle between the first and second orientations can be determined using the following formula, and the candidate target lane corresponding to the determined second orientation angle can be identified as the target lane:

[0073] fabs(Obs_end_heading-pos_heading)

[0074] Among them, Obs_end_heading can be used to represent the first orientation, pos_heading can be used to represent the second orientation, and fabs() can be used to represent the absolute value of the difference between the first orientation and the second orientation.

[0075] As an optional embodiment, the road coordinate system on the road segment includes a first coordinate axis and a second coordinate axis. Determining the target lane corresponding to the first predicted driving trajectory based on the trajectory points on the first predicted driving trajectory includes: determining the time it takes for the trajectory points on the first predicted driving trajectory to move to each of the at least one candidate target lanes along the first coordinate axis; determining the offset of the trajectory points on the first predicted driving trajectory when they move to the candidate target lanes along the second coordinate axis according to the time; and determining the target lane based on the offset.

[0076] In this embodiment, the road coordinate system on the road segment may include a first coordinate axis and a second coordinate axis. Determining the time it takes for a trajectory point on the first predicted driving trajectory to move to each of the at least one candidate target lanes along the first coordinate axis allows us to determine the offset of the trajectory point on the first predicted driving trajectory when it moves to the candidate target lane along the second coordinate axis, and the target lane can be determined based on this offset. The first coordinate axis may be the longitudinal coordinate axis (S) of the road. The second coordinate axis may be the transverse coordinate axis (L) of the road.

[0077] Optionally, the first coordinate axis and the second coordinate axis in the road coordinate system on the road segment can be determined, the time it takes for the trajectory point on the first predicted driving trajectory to move to each of the candidate target lanes in at least one candidate target lane according to the first coordinate axis can be determined, the arrival time can be obtained, the offset of the trajectory point on the first predicted driving trajectory when it moves to the candidate target lane according to the second coordinate axis can be determined based on the arrival time, and the target lane can be determined based on the offset.

[0078] For example, the arrival time of the autonomous vehicle in the longitudinal direction to the candidate target lane can be determined based on the projection of the autonomous vehicle's speed onto the longitudinal axis of the road (V_S) and the longitudinal coordinate of the road (S). The offset of the autonomous vehicle in the lateral coordinate system can be determined based on the arrival time and the projection of its speed onto the lateral axis of the road (V_L), and the target lane can be determined based on the offset.

[0079] As an optional embodiment, determining the target lane based on the offset includes: among at least one candidate target lane, determining the candidate target lane corresponding to the smallest offset as the target lane.

[0080] In this embodiment, the offset of each candidate target lane can be determined, and the candidate target lane with the smallest offset can be determined as the target lane from at least one candidate target lane.

[0081] In this embodiment of the disclosure, by establishing a road coordinate system, the time it takes for a trajectory point on the first predicted driving trajectory to move to each of the at least one candidate target lanes according to the first coordinate axis in the road coordinate system is determined; the offset of the trajectory point on the first predicted driving trajectory when it moves to the candidate target lane according to the second coordinate axis in the road coordinate system is determined according to the time; the candidate target lane with the smallest offset is determined as the target lane, thereby selecting the most suitable target lane from multiple candidate target lanes, which achieves the effect of improving the accuracy of path prediction.

[0082] For example, the arrival time of the autonomous vehicle in the longitudinal direction to the candidate target lane can be determined by projecting the vehicle's speed onto the longitudinal axis of the road. The offset of the autonomous vehicle on the lateral axis can be determined by projecting the arrival time and speed onto the lateral axis of the road, and the candidate target lane corresponding to the minimum offset can be identified as the target lane.

[0083] As an optional embodiment, the third orientation of the autonomous vehicle and the fourth orientation of each of the at least one candidate target lanes are obtained; in response to the angle between the third orientation and the fourth orientation satisfying an angle threshold, and the coordinate value of the autonomous vehicle moving on the second coordinate axis satisfying a coordinate threshold, the candidate target lane is determined as the target lane for enabling the autonomous vehicle to perform target steering.

[0084] In this embodiment, the third orientation of the autonomous vehicle and the fourth orientation of each of at least one candidate target lane can be obtained. It can be determined whether the angle between the third and fourth orientations meets an angle threshold. If the angle (Heading_diff) between the third and fourth orientations meets the angle threshold, and the coordinate value of the autonomous vehicle moving on the second coordinate axis meets the coordinate threshold, then the candidate target lane can be determined as a target lane for enabling the autonomous vehicle to perform a target turn. The third orientation can be the direction the autonomous vehicle is facing (current orientation). The fourth orientation can be the orientation of the candidate target lane (exit orientation). The angle threshold (also called the difference threshold) and the coordinate threshold (also called the lateral distance threshold) can be values ​​preset based on experience or actual conditions; no specific restrictions are placed on the method for determining the angle threshold and coordinate threshold here. The target turn can be a preset turn, such as a left turn or a right turn; this is only an example, and no specific restrictions are placed on the target turn. The coordinate value of the autonomous vehicle on the second coordinate axis can be the lateral distance (L) of the road autonomous driving on the second coordinate axis.

[0085] Optionally, for trajectories with relatively small curvature or dedicated left-turn lanes, a left turn may be the better choice based on the usage scenario. However, it is impossible to accurately determine whether to turn left solely based on the direction of the autonomous vehicle's front (third direction) and the direction of the candidate target lane (fourth direction). In this case, the embodiments of this disclosure determine the angle between the autonomous vehicle's third and fourth directions. In response to the angle between the third and fourth directions satisfying an angle threshold and the coordinate value of the autonomous vehicle moving on the second coordinate axis satisfying a coordinate threshold, the candidate target lane is determined as the target lane for the autonomous vehicle to make the target turn. This achieves the technical effect of improving the accuracy of road prediction and solves the technical problem in the prior art that it is impossible to accurately predict the road.

[0086] For example, the difference (heading_diff) between the current orientation of the autonomous vehicle's front and the exit orientation (Lane_heading) of the candidate target lane can be determined to obtain the angle between the third and fourth orientations. Whether to choose to turn left can be determined using the lateral distance (L) on the second coordinate axis and the heading_diff. Coordinate thresholds (lateral distance threshold), angle thresholds (difference threshold), and the target direction being a left turn can be preset. It can be determined whether the autonomous vehicle's coordinate value (lateral distance) on the second coordinate axis meets the coordinate threshold, and whether the angle between the third and fourth orientations meets the angle threshold. In response to the lateral distance meeting the coordinate threshold and the angle between the third and fourth orientations meeting the angle threshold, it can be determined that a left turn is required, and the candidate target lane on the left can be identified as the target lane for the autonomous vehicle to make the target turn; if the lateral distance does not meet the coordinate threshold, and / or the angle between the third and fourth orientations does not meet the angle threshold, it can be determined that the autonomous vehicle does not need to turn left.

[0087] As an optional embodiment, generating a second predicted driving trajectory for an autonomous vehicle based on trajectory points and a target lane includes: determining a target trajectory point at a second time on the target lane based on the trajectory points at a first time, wherein the second time is after the first time; and generating a second predicted driving trajectory based at least on the trajectory points at the first time and the target trajectory point.

[0088] In this embodiment, a trajectory point at a first moment and a target trajectory point on the target lane at a second moment can be determined. Based on the trajectory point at the first moment and the target trajectory point, a second predicted driving trajectory for the autonomous vehicle can be generated. The trajectory point at the first moment can be referred to as a constraint point. The first and second moments can be moments determined experimentally or empirically, such as 5 seconds or 8 seconds; this is merely an example, and no specific limitations are placed on the magnitude or determination method of the first and second moments. The second moment occurs after the first moment. The target trajectory point can be the target point of the target lane, also known as the lane center point.

[0089] For example, constraint points can be determined based on empirical values ​​to determine the trajectory point at the first moment. For instance, the trajectory point at 5 seconds (5S) can be determined as the trajectory point at the first moment. To determine the target trajectory point at the second moment after the first moment on the target lane, a second predicted driving trajectory can be determined based on the constraint points and the target trajectory point using a spline algorithm.

[0090] In this embodiment of the disclosure, the second predicted driving trajectory is determined by the constraint point at the first moment and the target trajectory point at the second moment. This avoids the technical problem in the prior art of directly predicting the predicted driving trajectory based on road information, which leads to low accuracy of trajectory prediction. This further improves the accuracy of trajectory prediction and solves the technical problem of low accuracy of trajectory prediction.

[0091] As an optional embodiment, determining the target trajectory point at the second moment on the target lane based on the trajectory point at the first moment includes: determining the target trajectory point based on the trajectory point at the first moment and the orientation of the target lane.

[0092] In this embodiment, the speed at which the autonomous vehicle reaches the target point can be determined based on the speed of the trajectory point at the first moment and the orientation of the target lane. If the speed at the target point exceeds the preset maximum speed, the vehicle can continue along the target lane to determine the target trajectory point, thereby achieving the purpose of determining the target trajectory point based on the trajectory point at the first moment and the orientation of the target lane.

[0093] As an optional embodiment, generating a second predicted driving trajectory based at least on the trajectory point and the target trajectory point at a first time step includes: generating a second predicted driving trajectory based on the trajectory point at a third time step, the trajectory point at the first time step, and the target trajectory point, wherein the third time step is prior to the first time step.

[0094] In this embodiment, a second predicted driving trajectory can be determined based on the trajectory point at the third time moment, the trajectory point at the first time moment, and the target trajectory point. The third time moment can be a time preceding the first time moment, or a time preset based on experience, such as 1 second, 3 seconds, etc. This is merely an example, and no specific limitations are placed on the size or determination method of the third time moment. The trajectory point at the third time moment can also be called a constraint point, which can be used to determine the second predicted driving trajectory.

[0095] Optionally, a third time point, a first time point, and a second time point can be preset, wherein the third time point precedes the first time point, and the first time point precedes the second time point. By determining the trajectory points at the third time point, the trajectory points at the first time point, and the target trajectory points at the second time point, a second predicted driving trajectory can be generated based on these points.

[0096] For example, the first, second, and third moments can be determined based on empirical values; for instance, trajectory points at 1 second, 3 seconds, and 5 seconds can be used as constraint points. The target trajectory point at the second moment can be determined on the target lane, and the trajectory line can be determined based on the constraint points and the target trajectory point.

[0097] As an optional embodiment, in response to the second predicted driving trajectory not conforming to the target driving rule, a trajectory point at a fourth time is determined on the second predicted driving trajectory, wherein the fourth time is prior to the second time; the second predicted driving trajectory is adjusted based at least on the trajectory point at the fourth time and the target trajectory point, wherein the adjusted second predicted driving trajectory conforms to the target driving rule.

[0098] In this embodiment, it can be determined whether the second predicted driving trajectory conforms to the target driving rules. If the second predicted driving trajectory does not conform to the target driving rules, a trajectory point at a fourth time step can be determined on the second predicted driving trajectory. The second predicted driving trajectory can be adjusted based at least on the trajectory point at the fourth time step and the target trajectory point to obtain a second predicted driving trajectory that conforms to the target driving rules. The fourth time step can be a time step prior to the second time step; for example, if the second time step is 8 seconds, the fourth time step can be 4 seconds.

[0099] Optionally, the second predicted driving trajectory can be evaluated to determine whether it conforms to the target driving rules. If the second predicted driving trajectory does not conform to the target driving rules, trajectory points that do not conform to the target driving rules can be deleted, and trajectory points located at the fourth time point before the second time point can be determined, thereby completing the backtracking of the trajectory line. Based on the trajectory points at the fourth time point and the target trajectory points, the second predicted driving trajectory can be adjusted to obtain a second predicted driving trajectory that conforms to the target driving rules.

[0100] For example, if the constraint point at the 5-second mark exceeds the road centerline, but the target driving rules prohibit trajectory points from exceeding the road centerline, then the second predicted driving trajectory at this time can be determined to be inconsistent with the target driving rules. Further checks can be made to determine if the trajectory point at the 4-second mark exceeds the road centerline, until a trajectory point that does not exceed the road centerline and whose orientation meets the conditions is found. This determines the trajectory point at the fourth moment. The second predicted driving trajectory can then be adjusted based on the trajectory point at the fourth moment to obtain a second predicted driving trajectory that conforms to the target driving rules.

[0101] In related technologies, only trajectory points are determined to establish a predicted driving trajectory. However, this method cannot process the predicted driving trajectory based on prior knowledge and map information. The predicted trajectory may not be suitable for normal driving, or there may be risks of traffic violations during driving, resulting in low accuracy in the predicted driving trajectory. To avoid these problems, this disclosure adjusts the trajectory points in the second predicted driving trajectory based on the target driving rules, thereby improving the accuracy of the predicted driving trajectory and avoiding risks such as the predicted driving trajectory not conforming to driving rules or traffic regulations.

[0102] As an optional embodiment, in response to the second predicted driving trajectory being a reverse trajectory, it is determined that the second predicted driving trajectory does not conform to the target driving rules, and the second predicted driving trajectory is deleted; in response to the autonomous vehicle being in an accelerating state and the longitudinal distance of the second predicted driving trajectory meeting a first longitudinal distance threshold, or in response to the autonomous vehicle being in a decelerating state and the longitudinal distance of the second predicted driving trajectory meeting a second longitudinal distance threshold, it is determined that the second predicted driving trajectory does not conform to the target driving rules, and the longitudinal distance of the second predicted driving trajectory is adjusted based on the acceleration of the autonomous vehicle, wherein the first longitudinal distance threshold is determined based on the current speed and acceleration of the autonomous vehicle, and the second longitudinal distance threshold is determined based on acceleration.

[0103] In this embodiment, it is determined whether the second predicted driving trajectory is a reverse trajectory. If the second predicted driving trajectory is a reverse trajectory, it can be determined that driving in the second predicted driving trajectory is not allowed. In response to the second predicted driving trajectory being a reverse trajectory, it can be determined that the second predicted driving trajectory does not conform to the target driving rules, and the second predicted driving trajectory can be deleted.

[0104] In this embodiment, the driving state of the autonomous vehicle can be determined. In response to the autonomous vehicle being in an accelerating state and the longitudinal distance of the second predicted driving trajectory meeting a first longitudinal distance threshold, or in response to the autonomous vehicle being in a decelerating state and the longitudinal distance of the second predicted driving trajectory meeting a second longitudinal distance threshold, it can be determined that the second predicted driving trajectory does not conform to the target driving rules. When the second predicted driving trajectory does not conform to the target driving rules, the longitudinal distance of the second predicted driving trajectory can be adjusted based on the acceleration of the autonomous vehicle. The first longitudinal threshold can be a preset value, or it can be the maximum longitudinal distance. The second longitudinal threshold can be a longitudinal distance threshold calculated based on acceleration.

[0105] In this embodiment, the determined second predicted driving trajectory is preprocessed to delete the reverse second predicted driving trajectory, and the longitudinal distance of the second predicted driving trajectory that does not conform to the target driving rules is adjusted, thereby obtaining the second predicted driving trajectory that conforms to the target driving rules, thereby achieving the effect of improving the accuracy and drivability of trajectory prediction.

[0106] For example, the reversed second predicted driving trajectory can be deleted; or the second predicted driving trajectory can be adjusted according to the target driving rules. Adjusting the second predicted driving trajectory according to the target driving rules can include: calculating the maximum longitudinal distance based on the current speed and maximum acceleration of the autonomous vehicle; when the longitudinal distance of the determined trajectory line (second predicted driving trajectory) is greater than the target driving rules, the shape of the second predicted driving trajectory can be kept unchanged, and the longitudinal distance of the second predicted driving trajectory can be adjusted based on the current acceleration of the autonomous vehicle to obtain the adjusted second predicted driving trajectory. Alternatively, when the autonomous vehicle decelerates for 5 consecutive frames, it is determined that the autonomous vehicle is in a deceleration state, and the longitudinal distance of the second predicted driving trajectory is less than the longitudinal distance threshold calculated based on acceleration. In response to the autonomous vehicle being in a deceleration state and the longitudinal distance of the second predicted driving trajectory meeting the second longitudinal distance threshold, it can be determined that the second predicted driving trajectory does not conform to the target driving rules. The shape of the second predicted driving trajectory can be kept unchanged, and the longitudinal distance of the trajectory line can be adjusted based on the current acceleration to obtain the adjusted second predicted driving trajectory.

[0107] As an optional embodiment, determining at least one first predicted driving trajectory that conforms to the target driving rules in at least one predicted driving trajectory includes: selecting a target number of predicted driving trajectories whose probability values ​​meet a probability threshold from at least one predicted driving trajectory; removing abnormal trajectory points from the target number of predicted driving trajectories, or removing abnormal predicted driving trajectories from the target number of predicted driving trajectories, to obtain at least one first predicted driving trajectory.

[0108] In this embodiment, among at least one predicted driving trajectory, the target number of predicted driving trajectories with probability values meeting the probability threshold can be selected. The abnormal trajectory points in the target number of predicted driving trajectories can be removed respectively, or the abnormal predicted driving trajectories can be removed from the target number of predicted driving trajectories, so as to obtain at least one first predicted driving trajectory. The probability threshold can be a value set in advance according to experience.

[0109] For example, the target number of predicted driving trajectories (for example, three predicted driving trajectories) can be selected, and the sum of the predicted driving trajectory probabilities is as close to 1 as possible. The multiple predicted driving trajectories (anchor trajectories) are sorted based on the probability values predicted by the predicted driving trajectories. When the probability value of the first predicted driving trajectory is greater than 0.5 and the probability value of the second predicted driving trajectory (top2) is less than 0.3 (Top1>0.5&&top2<0.3), the first predicted driving trajectory (TOP1) is output to obtain at least one first predicted driving trajectory. When the probability value of the first predicted driving trajectory is greater than 0.5 and the probability value of the second predicted driving trajectory is greater than 0.3 or the probability value of the first predicted driving trajectory is between 0.3 and 0.5 (Top1>0.5&&top2>0.3||0.3<Top1<0.5), the first predicted driving trajectory and the second predicted driving trajectory can be output to obtain two first predicted driving trajectories. When the probability value of the first predicted driving trajectory is less than 0.3, the top three predicted driving trajectories are output to obtain three first predicted driving trajectories. It should be noted that the number of the above first predicted driving trajectories is only for illustrative purposes, and the number of the first predicted driving trajectories is not specifically limited here. The unreasonable trajectory points can be determined according to prior knowledge and map information, and the unreasonable trajectory points and / or trajectory lines can be deleted.

[0110] In the embodiment of the present disclosure, from at least one predicted driving trajectory of the autonomous vehicle, a reasonable first predicted driving trajectory that conforms to the target driving rules is determined, and the corresponding target lane (exit lane) is determined based on the reasonable trajectory points on the first predicted driving trajectory, so as to generate a second predicted driving trajectory that conforms to the target driving rules based on the reasonable trajectory points and the target lane, thereby achieving the technical effect of effectively generating the driving trajectory of the autonomous vehicle and solving the technical problem of being unable to effectively generate the driving trajectory of the autonomous vehicle.

[0111] The embodiment of the present disclosure also provides another method for generating a predicted driving trajectory of an autonomous vehicle from the human-computer interaction side. Figure 2a It is a flowchart of another method for generating a predicted driving trajectory of an autonomous vehicle according to the embodiment of the present disclosure, as Figure 2aAs shown, an implementation scheme for the method of generating the predicted driving trajectory of the autonomous vehicle may include at least the following implementation steps:

[0112] Step S202: Display at least one predicted driving trajectory output by the trajectory prediction model on the operation interface, wherein the trajectory prediction model is used to predict the driving trajectory of the autonomous vehicle.

[0113] In the technical solution provided by step S202 of this application, the trajectory prediction model predicts the driving trajectory of the autonomous vehicle, obtaining at least one predicted driving trajectory, which can be displayed on the operation interface. The trajectory prediction model can be used to predict the driving trajectory of the autonomous vehicle. It should be noted that any model capable of predicting the driving trajectory of an autonomous vehicle should be within the protection scope of this disclosure; no specific limitations are made on the type of trajectory prediction model here. The operation interface can be the display interface of the planning module, the interface of a mobile terminal, etc.; no specific limitations are made on the operation interface here.

[0114] Step S204: Display at least one first predicted driving trajectory that conforms to the target driving rule among at least one predicted driving trajectory on the operation interface, wherein the target driving rule is used to guide the autonomous vehicle to drive in a normal driving state on the road segment.

[0115] In the technical solution provided in step S204 of this disclosure, at least one first predicted driving trajectory that conforms to the target driving rule can be displayed on the operation interface. The target driving rule can be prior knowledge and / or map information, which can be used to filter the at least one predicted driving trajectory and guide the autonomous vehicle to drive in a normal driving state on the road segment. A normal driving state can refer to a driving state without violations of traffic regulations such as illegal driving or driving against traffic; this is only an example and does not impose specific limitations on the normal driving state. The first predicted driving trajectory can be a reasonable trajectory line among the at least one predicted driving trajectory.

[0116] Optionally, a target driving rule can be predetermined, and at least one first predicted driving trajectory that conforms to the target driving rule can be determined from at least one predicted driving trajectory. The at least one first predicted driving trajectory can be displayed on the operation interface.

[0117] Step S206: In response to the trajectory generation command applied to the operation interface, at least one first predicted driving trajectory is displayed on the operation interface as a second predicted driving trajectory. The second predicted driving trajectory is used to control the autonomous vehicle to drive to the target lane and is generated based on the trajectory points on the first predicted driving trajectory and the target lane. The target lane is the lane that the autonomous vehicle is to drive to and that conforms to the target driving rules. The target lane is determined on the road segment based on the trajectory points on the first predicted driving trajectory.

[0118] In the technical solution provided by step S206 of this application, a trajectory generation instruction on the operation interface can be obtained. In response to the obtained trajectory generation instruction, a second predicted driving trajectory from at least one first predicted driving trajectory can be displayed on the operation interface. The second predicted driving trajectory can be a driving trajectory that conforms to the target driving rules and can be used to control the autonomous vehicle to drive into the target lane. The trajectory generation instruction can be a spatially input instruction through the operation interface and can be used to indicate the second predicted driving trajectory determined from at least one first predicted driving trajectory.

[0119] For example, in response to a trajectory generation command applied to the user interface, the orientation of the last trajectory point can be determined based on the trajectory points on the first predicted driving trajectory. Based on the orientation of the last trajectory point, a target lane that conforms to the target driving rules corresponding to the first predicted driving trajectory can be determined on the road segment. Based on the trajectory points and the target lane, a second predicted driving trajectory for the autonomous vehicle can be generated. At least one second predicted driving trajectory from the first predicted driving trajectory can be displayed on the user interface.

[0120] Figure 2b This is a schematic diagram of an operation interface according to an embodiment of the present disclosure, such as... Figure 2b As shown, at least one predicted driving trajectory output by the trajectory prediction model can be displayed on the operation interface; at least one first predicted driving trajectory that conforms to the target driving rules can be displayed among the at least one predicted driving trajectory on the operation interface; a trajectory generation command can be issued by clicking the trajectory generation control in the operation interface. In response to the trajectory generation command applied to the operation interface, a second predicted driving trajectory among the at least one first predicted driving trajectory can be displayed on the operation interface.

[0121] Through steps S202 to S206 above, at least one predicted driving trajectory output by the trajectory prediction model is displayed on the operation interface, wherein the trajectory prediction model is used to predict the driving trajectory of the autonomous vehicle; at least one first predicted driving trajectory that conforms to the target driving rules is displayed on the operation interface, wherein the target driving rules are used to guide the autonomous vehicle to drive in a normal driving state on the road segment; in response to the trajectory generation command applied to the operation interface, a second predicted driving trajectory is displayed on the operation interface, wherein the second predicted driving trajectory is used to control the autonomous vehicle to drive to the target lane, and is generated based on the trajectory points on the first predicted driving trajectory and the target lane, wherein the target lane is the lane that the autonomous vehicle is to drive to and conforms to the target driving rules, and the target lane is determined on the road segment based on the trajectory points on the first predicted driving trajectory. In other words, in this embodiment of the present disclosure, a reasonable first predicted driving trajectory that conforms to the target driving rules is determined from at least one predicted driving trajectory of the autonomous vehicle, and the corresponding target lane (exit lane) is determined based on the reasonable trajectory points on the first predicted driving trajectory. A second predicted driving trajectory that conforms to the target driving rules is then generated based on the reasonable trajectory points and the target lane, thereby achieving the technical effect of effectively generating the driving trajectory of the autonomous vehicle and solving the technical problem of not being able to effectively generate the driving trajectory of the autonomous vehicle.

[0122] The above technical solutions of the present disclosure will be further illustrated below with reference to preferred embodiments.

[0123] Predicting the driving trajectory has a significant impact on downstream planning modules. A reasonable and accurate prediction of the driving trajectory can better improve the intelligence and driving rationality of autonomous vehicles.

[0124] Currently, predicted driving trajectories are typically generated by outputting continuous points from the model. However, the model has a large number of inputs, mostly raw perception information, which contains some implicit knowledge that the model cannot learn. As a result, the output trajectory points cannot guarantee that all expert prior information is taken into account, nor can they perfectly match the current map information. This often leads to obviously erroneous trajectories. At the same time, the output discrete trajectories cannot meet downstream requirements. For example, when downstream applications require a smooth and continuous prediction of the driving trajectory for 8 seconds, related technologies intelligently output several discrete trajectory points with large intervals.

[0125] To solve the above problems, the embodiments of the present disclosure combine discrete prediction model points with prior driving information and map information, perform a series of operations such as deletion and adjustment on trajectory points, and finally use the spline algorithm to generate a smooth curve with constraint points, meeting the requirements of the autonomous vehicle for the predicted driving trajectory. The embodiments of the present disclosure process the predicted trajectory based on prior driving information and map information, solving the technical problem in the prior art that only a predicted driving trajectory is generated through a trajectory prediction model, resulting in a low accuracy of the generated predicted driving trajectory.

[0126] Figure 3 is a flowchart of a driving trajectory generation method according to an embodiment of the present disclosure. As Figure 3 shown, the driving trajectory generation method that combines the output of the trajectory prediction model with map information and expert prior information may include the following steps.

[0127] Step S302, select reasonable trajectory points.

[0128] In this embodiment, the trajectory prediction model can predict multiple predicted driving trajectories, and the possible driving trajectory of the autonomous vehicle can be determined according to the probability value of the predicted driving trajectory. Among them, the probability value can be determined according to actual experience or calculated through multiple experiments. The specific method for determining the probability value is not limited herein.

[0129] For example, a target number of predicted driving trajectories (such as three predicted driving trajectories) can be selected, and the sum of the probabilities of the predicted driving trajectories is as close to 1 as possible. Sort the multiple predicted driving trajectories based on the predicted probability value. When the probability value of the first predicted driving trajectory is greater than 0.5 and the probability value of the second predicted driving trajectory is less than 0.3, the first predicted driving trajectory can be output. When the probability value of the first predicted driving trajectory is greater than 0.5 and the probability value of the second predicted driving trajectory is greater than 0.3 or the probability value of the first predicted driving trajectory is between 0.3 and 0.5 (Top1>0.5&&top2>0.3||0.3<Top1<0.5), the first predicted driving trajectory and the second predicted driving trajectory can be output. When the probability value of the first predicted driving trajectory is less than 0.3, the first, second, and third predicted driving trajectories ranked top three can be output.

[0130] Step S304, delete unreasonable trajectory points and / or trajectory lines.

[0131] In this embodiment, unreasonable trajectory points can be determined according to prior knowledge and map information, and the unreasonable trajectory points and / or trajectory lines can be deleted.

[0132] For example, Figure 4This is a schematic diagram of a trajectory point outside an intersection according to an embodiment of this disclosure, such as... Figure 4 As shown, when a trajectory point is outside an intersection, it can be determined based on prior knowledge and / or map information that the autonomous vehicle cannot leave such a trajectory. Therefore, the trajectory point can be deleted. Figure 5 This is a schematic diagram of a trajectory point crossing a boundary according to an embodiment of this disclosure, such as... Figure 5 As shown, when the predicted driving trajectory crosses an insurmountable boundary or a yellow line, it can be determined based on prior knowledge and / or map information that the autonomous vehicle cannot leave such a trajectory, and therefore the predicted driving trajectory can be deleted. Figure 6 This is a schematic diagram of trajectory point reversal according to an embodiment of the present disclosure, such as... Figure 6 As shown, when the predicted trajectory points regress, it can be determined based on prior knowledge and / or map information that the autonomous vehicle cannot leave such a trajectory, and therefore the predicted driving trajectory can be deleted.

[0133] Step S306: Determine the exit lane based on the trajectory points.

[0134] In this embodiment, Figure 7 This is a schematic diagram of a target lane selection according to an embodiment of the present disclosure, such as... Figure 7 As shown, it can be determined whether the last trajectory point of the predicted driving trajectory is on the candidate target lane (which can be the actual exit lane). If the last trajectory point is on the candidate target lane, the candidate target lane can be determined as the target lane. If the last trajectory point is not on the exit lane, the first orientation can be determined based on the last two trajectory points of the predicted driving trajectory, and the second orientation from the last trajectory point to the candidate target lane can be determined based on the first and second orientations.

[0135] Optionally, the first orientation can be determined by the orientation of the last two trajectory points, and the second orientation can be determined by the orientation of the last trajectory point towards the candidate target lane. When the last trajectory point is not on the exit lane, the angle between the first orientation and each second orientation can be determined separately, and the second orientation with the smallest angle to the first orientation can be selected. The candidate target lane corresponding to the determined second orientation is then selected as the target lane. The difference between the angles of the first orientation and each second orientation can be determined using the following formula:

[0136] fabs(Obs_end_heading-pos_heading)

[0137] Here, `Obs_end_heading` can be used to represent the first orientation, `pos_heading` can be used to represent the second orientation, and `fabs()` can be used to represent the absolute value of the difference between the first and second orientations.

[0138] In this embodiment, for some predicted driving trajectories with small curvature, or on dedicated left-turn lanes, a better choice for left turns can be determined based on scenario reasoning, but it is difficult to make a judgment based solely on heading_diff.

[0139] Figure 8 This is a schematic diagram of the orientation of an autonomous vehicle according to an embodiment of the present disclosure, such as... Figure 8 As shown, the difference between the current orientation of the vehicle and the exit orientation of the candidate target lane can be determined first. Then, using the lateral distance (L) and heading_diff on the second coordinate axis, a decision can be made on whether to turn left. For example, a lateral distance threshold and a difference threshold can be preset. The system checks whether the lateral distance and the difference meet the thresholds. If both the lateral distance and the difference meet the thresholds, a left turn is required. If neither the lateral distance nor the difference meets the threshold, a left turn is not required.

[0140] Figure 9 This is a schematic diagram of a target lane selection according to an embodiment of the present disclosure, such as... Figure 9 As shown, candidate exit lanes can be used as reference lanes. The arrival time of the autonomous vehicle in the longitudinal direction to the reference lane can be determined based on the projection of speed on the longitudinal axis of the road (V_S) and the longitudinal coordinate of the road (S). The offset of the autonomous vehicle in the lateral coordinate system can be determined based on the arrival time and the projection of speed on the lateral axis of the road (V_L). The reference lane with the smallest offset can be determined as the target lane.

[0141] Step S308: Generate a trajectory line based on the trajectory points and the exit lane.

[0142] In this embodiment, the trajectory points and the target points of the exit lane can be processed using a spline algorithm to obtain the trajectory line.

[0143] For example, Figure 10 This is a schematic diagram illustrating the selection of constraint points according to an embodiment of the present disclosure, such as... Figure 10As shown, constraint points can be determined based on empirical values. For example, trajectory points at 1 second (1s), 3 seconds (3s), and 5 seconds (5s) can be used as constraint points. A target point is determined on the target lane, and the trajectory line is determined based on the constraint points and the target point of the target lane. The target point of the target lane can be the center point of the target lane. The speed at which the autonomous vehicle reaches the target point can be determined based on the speed of the last trajectory point and the orientation of the target lane. If the speed at the target point is too high, the constraint points can be backtracked. Figure 11a This is a schematic diagram of constraint point backtracking according to an embodiment of the present disclosure, as shown below. Figure 11a As shown, the constraint points can be traced back. For example, if the constraint point at the 5-second position is found to be beyond the center of the road, the trajectory point at the 4-second position can be further determined to see if it exceeds the center line of the road, until the constraint point found does not exceed the center line of the road and is oriented towards a constraint point that meets the conditions.

[0144] Optionally, Figure 11b This is a schematic diagram of the generation result of a trajectory line according to an embodiment of the present disclosure, such as... Figure 11b As shown, a trajectory line can be determined based on defined constraint points and target points, and autonomous vehicles can be controlled to drive according to the trajectory line.

[0145] Step S310: Post-process the trajectory line.

[0146] In this embodiment, Figure 11c This is a schematic diagram of the post-processing result of a trajectory line according to an embodiment of the present disclosure, such as... Figure 11c As shown, to further improve the accuracy of trajectory prediction, post-processing methods can be used to remove abnormal trajectories, such as deleting inverse trajectories; or the trajectory can be adjusted according to actual road conditions to obtain the final predicted driving trajectory, such as... Figure 11c The position indicated by the middle arrow.

[0147] Optionally, the maximum longitudinal distance is calculated based on the current speed and maximum acceleration of the autonomous vehicle. When the longitudinal distance of the determined trajectory line is greater than the maximum longitudinal distance, the shape of the trajectory line can remain unchanged. The longitudinal distance of the trajectory line is adjusted based on the current acceleration to obtain the adjusted trajectory line.

[0148] Optionally, when the autonomous vehicle decelerates for 5 consecutive frames and the longitudinal distance of the trajectory line is less than the longitudinal distance threshold calculated based on acceleration, the shape of the trajectory line can remain unchanged, and the longitudinal distance of the trajectory line can be adjusted based on the current acceleration to obtain the adjusted trajectory line.

[0149] In this embodiment of the disclosure, a reasonable first predicted driving trajectory that conforms to the target driving rules is determined from at least one predicted driving trajectory of the autonomous vehicle. The corresponding target lane (exit lane) is determined based on the reasonable trajectory points on the first predicted driving trajectory. A second predicted driving trajectory that conforms to the target driving rules is generated based on the reasonable trajectory points and the target lane. This achieves the technical effect of effectively generating the driving trajectory of the autonomous vehicle and solves the technical problem of not being able to effectively generate the driving trajectory of the autonomous vehicle.

[0150] This disclosure also provides an embodiment for performing Figure 1 The illustrated embodiment is a method for generating a predicted driving trajectory for an autonomous vehicle and an apparatus for generating a predicted driving trajectory for an autonomous vehicle.

[0151] Figure 12 This is a schematic diagram of a predictive driving trajectory generation device for an autonomous vehicle according to an embodiment of the present disclosure. Figure 12 As shown, the predictive driving trajectory generation device 1200 for the autonomous vehicle may include: an acquisition unit 1202, a first determination unit 1204, a second determination unit 1206, and a generation unit 1208.

[0152] The acquisition unit 1202 is used to acquire at least one predicted driving trajectory of the autonomous vehicle.

[0153] The first determining unit 1204 is used to determine at least one first predicted driving trajectory that conforms to the target driving rule from at least one predicted driving trajectory, wherein the target driving rule is used to guide the autonomous vehicle to drive in a normal driving state on the road segment.

[0154] The second determining unit 1206 is used to determine the target lane corresponding to the first predicted driving trajectory on a road segment based on the trajectory points on the first predicted driving trajectory, wherein the target lane is the lane that the autonomous vehicle is to drive to and that conforms to the target driving rules.

[0155] The generation unit 1208 is used to generate a second predicted driving trajectory for an autonomous vehicle based on trajectory points and a target lane, wherein the second predicted driving trajectory is used to control the autonomous vehicle to drive to the target lane.

[0156] Optionally, the first determining unit 1204 includes: a first determining module, used to determine, in at least one predicted driving trajectory, the predicted driving trajectory that matches the map information of the road segment as the first predicted driving trajectory.

[0157] Optionally, the first determining unit 1204 includes: a second determining module, used to determine the predicted driving trajectory that matches prior driving information as the first predicted driving trajectory in at least one predicted driving trajectory.

[0158] Optionally, the second determining unit 1206 includes: a third determining module, used to determine a target lane based at least on the last trajectory point on the first predicted driving trajectory, wherein multiple trajectory points on the first predicted driving trajectory are arranged according to their corresponding trajectory directions.

[0159] Optionally, the third determining module includes: a first determining submodule, configured to determine the candidate target lane as the target lane in response to the last trajectory point being on the candidate target lane; or, in response to the last trajectory point not being on the candidate target lane, to determine the target lane based on a first orientation determined by the last two trajectory points of the first predicted driving trajectory, and a second orientation of the last trajectory point pointing to each of the at least one candidate target lane.

[0160] Optionally, the third determining module further includes: a second determining submodule, used to determine, among at least one candidate target lane, a second orientation with the smallest angle to the first orientation; and to determine the candidate target lane corresponding to the determined second orientation as the target lane.

[0161] Optionally, the second determining unit 1208 includes: a first processing module, configured to determine the time it takes for a trajectory point on the first predicted driving trajectory to move to each of the at least one candidate target lanes along a first coordinate axis; determine the offset of the trajectory point on the first predicted driving trajectory when it moves to the candidate target lane along a second coordinate axis according to the time; and determine the target lane based on the offset.

[0162] Optionally, the processing module further includes: a second determining subunit, used to determine the candidate target lane corresponding to the minimum offset as the target lane among at least one candidate target lane.

[0163] Optionally, the processing module further includes: a first processing submodule, configured to acquire the third orientation of the autonomous vehicle and the fourth orientation of each of the at least one candidate target lanes; in response to the angle between the third orientation and the fourth orientation satisfying an angle threshold, and the coordinate value of the autonomous vehicle moving on the second coordinate axis satisfying a coordinate threshold, the candidate target lane is determined as the target lane for enabling the autonomous vehicle to perform target steering.

[0164] Optionally, the generation unit 1208 includes: a second processing module, configured to determine a target trajectory point at a second time on the target lane based on the trajectory point at the first time, wherein the second time is after the first time; and to generate a second predicted driving trajectory based at least on the trajectory point at the first time and the target trajectory point.

[0165] Optionally, the second processing module includes a third determining subunit, used to determine the target trajectory point based on the trajectory point at the first moment and the orientation of the target lane.

[0166] Optionally, the second processing module includes a generation subunit for generating a second predicted driving trajectory based on the trajectory point at the third time, the trajectory point at the first time, and the target trajectory point, wherein the third time is prior to the first time.

[0167] Optionally, the second processing module further includes: a second processing submodule, configured to, in response to the second predicted driving trajectory not conforming to the target driving rule, determine a trajectory point at a fourth time on the second predicted driving trajectory, wherein the fourth time is prior to the second time; and adjust the second predicted driving trajectory based at least on the trajectory point at the fourth time and the target trajectory point, wherein the adjusted second predicted driving trajectory conforms to the target driving rule.

[0168] Optionally, the apparatus further includes: a processing unit, configured to: in response to the second predicted driving trajectory being a reverse driving trajectory, determine that the second predicted driving trajectory does not conform to the target driving rules, and delete the second predicted driving trajectory; in response to the autonomous vehicle being in an accelerating state and the longitudinal distance of the second predicted driving trajectory meeting a first longitudinal distance threshold, or in response to the autonomous vehicle being in a decelerating state and the longitudinal distance of the second predicted driving trajectory meeting a second longitudinal distance threshold, determine that the second predicted driving trajectory does not conform to the target driving rules, and adjust the longitudinal distance of the second predicted driving trajectory based on the acceleration of the autonomous vehicle, wherein the first longitudinal distance threshold is determined based on the current speed and acceleration of the autonomous vehicle, and the second longitudinal distance threshold is determined based on the acceleration.

[0169] This disclosure also provides another device for generating a predicted driving trajectory of an autonomous vehicle for performing the predicted driving trajectory generation method of the autonomous vehicle shown in FIG2.

[0170] Figure 13 This is a schematic diagram of another predictive driving trajectory generation device for an autonomous vehicle according to an embodiment of the present disclosure. Figure 13 As shown, the predictive driving trajectory generation device 1300 for the autonomous vehicle may include: a first display unit 1302, a second display unit 1304 and a third display unit 1306.

[0171] The first display unit 1302 is used to display at least one predicted driving trajectory output by the trajectory prediction model on the operation interface, wherein the trajectory prediction model is used to predict the driving trajectory of the autonomous vehicle.

[0172] The second display unit 1304 is used to display on the operation interface at least one first predicted driving trajectory that conforms to the target driving rule in at least one predicted driving trajectory, wherein the target driving rule is used to guide the autonomous vehicle to drive in a normal driving state on the road segment.

[0173] The third display unit 1306 is used to respond to a trajectory generation command applied to the operation interface and display a second predicted driving trajectory in at least one first predicted driving trajectory on the operation interface. The second predicted driving trajectory is used to control the autonomous vehicle to drive to the target lane and is generated based on the trajectory points on the first predicted driving trajectory and the target lane. The target lane is the lane that the autonomous vehicle is to drive to and that conforms to the target driving rules. The target lane is determined on the road segment based on the trajectory points on the first predicted driving trajectory.

[0174] In the autonomous vehicle predictive driving trajectory generation device of the present disclosure embodiment, a reasonable first predicted driving trajectory that conforms to the target driving rules is determined from at least one predicted driving trajectory of the autonomous vehicle, and a corresponding target lane (exit lane) is determined based on the reasonable trajectory points on the first predicted driving trajectory. A second predicted driving trajectory that conforms to the target driving rules is generated based on the reasonable trajectory points and the target lane, thereby achieving the technical effect of effectively generating the driving trajectory of the autonomous vehicle and solving the technical problem of not being able to effectively generate the driving trajectory of the autonomous vehicle.

[0175] The acquisition, storage, and application of user personal information involved in the technical solution disclosed herein comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0176] According to embodiments of this disclosure, this disclosure also provides an electronic device, a readable storage medium, a computer program product, and an autonomous vehicle.

[0177] Embodiments of this disclosure provide an electronic device that may include: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the method for generating a predicted driving trajectory of an autonomous vehicle according to embodiments of this disclosure.

[0178] Optionally, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor and the input / output device is connected to the processor.

[0179] According to embodiments of this disclosure, an autonomous driving vehicle is also provided, wherein the autonomous driving vehicle may include the aforementioned electronic equipment.

[0180] According to embodiments of this disclosure, this disclosure also provides a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause a computer to execute the method for generating a predicted driving trajectory of an autonomous vehicle according to embodiments of this disclosure.

[0181] Optionally, in this embodiment, the non-volatile storage medium described above can be configured to store a computer program for performing the following steps:

[0182] S1, Obtain at least one predicted driving trajectory of the autonomous vehicle;

[0183] S2, in at least one predicted driving trajectory, determine at least one first predicted driving trajectory that conforms to the target driving rule, wherein the target driving rule is used to guide the autonomous vehicle to drive in a normal driving state on the road segment;

[0184] S3, based on the trajectory points on the first predicted driving trajectory, determine the target lane corresponding to the first predicted driving trajectory on the road segment, wherein the target lane is the lane that the autonomous vehicle is to drive to and that conforms to the target driving rules;

[0185] S4, based on the trajectory points and the target lane, generates a second predicted driving trajectory for the autonomous vehicle, wherein the second predicted driving trajectory is used to control the autonomous vehicle to drive to the target lane.

[0186] Optionally, in this embodiment, the non-volatile storage medium may also be configured to store a computer program for performing the following steps:

[0187] S1, Display at least one predicted driving trajectory output by the trajectory prediction model on the operation interface, wherein the trajectory prediction model is used to predict the driving trajectory of the autonomous vehicle.

[0188] S2, Display at least one first predicted driving trajectory that conforms to the target driving rule in at least one predicted driving trajectory on the operation interface, wherein the target driving rule is used to guide the autonomous vehicle to drive in a normal driving state on the road segment;

[0189] S3, responding to the trajectory generation command applied to the operation interface, displays at least one second predicted driving trajectory in a first predicted driving trajectory on the operation interface, wherein the second predicted driving trajectory is used to control the autonomous vehicle to drive to the target lane, and is generated based on the trajectory points on the first predicted driving trajectory and the target lane, the target lane is the lane to which the autonomous vehicle is to drive and which conforms to the target driving rules, and the target lane is determined on the road segment based on the trajectory points on the first predicted driving trajectory.

[0190] Optionally, in this embodiment, the aforementioned non-transitory computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or equipment, or any suitable combination of the foregoing. More specific examples of readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0191] According to embodiments of this disclosure, this disclosure also provides a computer program product, including a computer program that, when executed by a processor, performs the following steps:

[0192] S1, Obtain at least one predicted driving trajectory of the autonomous vehicle;

[0193] S2, in at least one predicted driving trajectory, determine at least one first predicted driving trajectory that conforms to the target driving rule, wherein the target driving rule is used to guide the autonomous vehicle to drive in a normal driving state on the road segment;

[0194] S3, based on the trajectory points on the first predicted driving trajectory, determine the target lane corresponding to the first predicted driving trajectory on the road segment, wherein the target lane is the lane that the autonomous vehicle is to drive to and that conforms to the target driving rules;

[0195] S4, based on the trajectory points and the target lane, generates a second predicted driving trajectory for the autonomous vehicle, wherein the second predicted driving trajectory is used to control the autonomous vehicle to drive to the target lane.

[0196] Optionally, in this embodiment, the computer program, when executed by the processor, may further perform the following steps:

[0197] S1, Display at least one predicted driving trajectory output by the trajectory prediction model on the operation interface, wherein the trajectory prediction model is used to predict the driving trajectory of the autonomous vehicle.

[0198] S2, Display at least one first predicted driving trajectory that conforms to the target driving rule in at least one predicted driving trajectory on the operation interface, wherein the target driving rule is used to guide the autonomous vehicle to drive in a normal driving state on the road segment;

[0199] S3, responding to the trajectory generation command applied to the operation interface, displays at least one second predicted driving trajectory in a first predicted driving trajectory on the operation interface, wherein the second predicted driving trajectory is used to control the autonomous vehicle to drive to the target lane, and is generated based on the trajectory points on the first predicted driving trajectory and the target lane, the target lane is the lane to which the autonomous vehicle is to drive and which conforms to the target driving rules, and the target lane is determined on the road segment based on the trajectory points on the first predicted driving trajectory.

[0200] Figure 14 This is a block diagram of an electronic device for generating a predicted driving trajectory for an autonomous vehicle according to an embodiment of the present disclosure. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0201] like Figure 14 As shown, device 1400 includes a computing unit 1401, which can perform various appropriate actions and processes according to a computer program stored in read-only memory (ROM) 1402 or a computer program loaded from storage unit 1408 into random access memory (RAM) 1403. RAM 1403 may also store various programs and data required for the operation of device 1400. The computing unit 1401, ROM 1402, and RAM 1403 are interconnected via bus 1404. Input / output (I / O) interface 1405 is also connected to bus 1404.

[0202] Multiple components in device 1400 are connected to I / O interface 1405, including: input unit 1406, such as keyboard, mouse, etc.; output unit 1404, such as various types of monitors, speakers, etc.; storage unit 1408, such as disk, optical disk, etc.; and communication unit 1409, such as network card, modem, wireless transceiver, etc. Communication unit 1409 allows device 1400 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0203] The computing unit 1401 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 1401 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 1401 performs the various methods and processes described above, such as the driving trajectory method for an autonomous vehicle. For example, in some embodiments, the method for generating a predicted driving trajectory for an autonomous vehicle can be implemented as a computer software program tangibly contained in a machine-readable medium, such as storage unit 1408. In some embodiments, part or all of the computer program can be loaded and / or installed on device 1400 via ROM 1402 and / or communication unit 1409. When the computer program is loaded into RAM 1403 and executed by the computing unit 1401, one or more steps of the data processing methods described above can be performed. Alternatively, in other embodiments, the computing unit 1401 may be configured by any other suitable means (e.g., by means of firmware) to perform a method for generating a predicted driving trajectory for an autonomous vehicle.

[0204] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0205] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0206] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0207] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display, monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0208] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with embodiments of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0209] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact via communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other. Servers can be cloud servers, servers in distributed systems, or servers incorporating blockchain technology.

[0210] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this disclosure can be achieved, and this is not limited herein.

[0211] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A method for generating a predicted driving trajectory for an autonomous vehicle, comprising: Obtain at least one predicted driving trajectory for an autonomous vehicle; Among the at least one predicted driving trajectory, at least one first predicted driving trajectory that conforms to the target driving rule is determined, wherein the target driving rule is used to guide the autonomous vehicle to drive in a normal driving state on the road segment; Based on the trajectory points on the first predicted driving trajectory, a target lane corresponding to the first predicted driving trajectory is determined on the road segment, wherein the target lane is the lane that the autonomous vehicle is to drive to and that conforms to the target driving rules; Based on the trajectory points and the target lane, a second predicted driving trajectory of the autonomous vehicle is generated, wherein the second predicted driving trajectory is used to control the autonomous vehicle to drive to the target lane; The road coordinate system on the road segment includes a first coordinate axis and a second coordinate axis. Determining the target lane corresponding to the first predicted driving trajectory based on the trajectory points on the first predicted driving trajectory includes: determining the time it takes for the trajectory points on the first predicted driving trajectory to move to each of the candidate target lanes according to the first coordinate axis; determining the offset of the trajectory points on the first predicted driving trajectory when they move to the candidate target lanes according to the second axis according to the time; and determining the target lane based on the offset.

2. The method according to claim 1, wherein, The target driving rule is determined by the map information of the road segment. Among the at least one predicted driving trajectory, the at least one first predicted driving trajectory that conforms to the target driving rule includes: Among the at least one predicted driving trajectory, the predicted driving trajectory that matches the map information of the road segment is determined as the first predicted driving trajectory.

3. The method according to claim 1, wherein, The target driving rule is determined by the prior driving information of the autonomous vehicle. Among the at least one predicted driving trajectory, determining at least one first predicted driving trajectory that conforms to the target driving rule includes: Among the at least one predicted driving trajectory, the predicted driving trajectory that matches the prior driving information is determined as the first predicted driving trajectory.

4. The method according to claim 1, wherein, Based on the trajectory points on the first predicted driving trajectory, the target lane corresponding to the first predicted driving trajectory is determined as follows: The target lane is determined based at least on the last trajectory point on the first predicted driving trajectory, wherein the multiple trajectory points on the first predicted driving trajectory are arranged according to the corresponding trajectory directions.

5. The method according to claim 4, wherein, Determining the target lane, based at least on the end trajectory point on the first predicted driving trajectory, includes: In response to the last trajectory point being located on the candidate target lane, the candidate target lane is determined as the target lane; or... In response to the last trajectory point not being in the candidate target lane, the target lane is determined based on a first orientation determined by the last two trajectory points of the first predicted driving trajectory, and a second orientation of the last trajectory point pointing to each of the candidate target lanes.

6. The method according to claim 5, wherein, The target lane is determined based on a first orientation determined by the last two trajectory points of the first predicted driving trajectory, and a second orientation pointing from the last trajectory point to each of the at least one candidate target lane. Among the at least one candidate target lane, determine the second orientation that has the smallest angle with the first orientation; The candidate target lane corresponding to the determined second orientation is determined as the target lane.

7. The method according to claim 1, wherein, Determining the target lane based on the offset includes: Among the at least one candidate target lane, the candidate target lane corresponding to the smallest offset is determined as the target lane.

8. The method according to claim 1, wherein, Also includes: Obtain the third orientation of the autonomous vehicle and the fourth orientation of each of the candidate target lanes in at least one candidate target lane; In response to the angle between the third orientation and the fourth orientation satisfying an angle threshold, and the coordinate value of the autonomous vehicle moving on the second coordinate axis satisfying a coordinate threshold, the candidate target lane is determined as the target lane for enabling the autonomous vehicle to perform a target turn.

9. The method according to claim 1, wherein, Based on the trajectory points and the target lane, generating the second predicted driving trajectory of the autonomous vehicle includes: Based on the trajectory points at the first moment, a target trajectory point at a second moment is determined on the target lane, wherein the second moment is after the first moment; The second predicted driving trajectory is generated based at least on the trajectory point and the target trajectory point at the first moment.

10. The method according to claim 9, wherein, Determining the target trajectory point at the second time on the target lane based on the trajectory point at the first time includes: The target trajectory point is determined based on the trajectory point at the first moment and the orientation of the target lane.

11. The method according to claim 9, wherein, Generating the second predicted driving trajectory, based at least on the trajectory point and the target trajectory point at the first time moment, includes: The second predicted driving trajectory is generated based on the trajectory point at the third time point, the trajectory point at the first time point, and the target trajectory point, wherein the third time point is prior to the first time point.

12. The method of claim 11, further comprising: In response to the second predicted driving trajectory not conforming to the target driving rule, the trajectory point at a fourth time point is determined on the second predicted driving trajectory, wherein the fourth time point is prior to the second time point; The second predicted driving trajectory is adjusted based at least on the trajectory point and the target trajectory point at the fourth time point, wherein the adjusted second predicted driving trajectory conforms to the target driving rule.

13. The method according to claim 1, further comprising: In response to the fact that the second predicted driving trajectory is a reverse driving trajectory, it is determined that the second predicted driving trajectory does not conform to the target driving rule, and the second predicted driving trajectory is deleted; In response to the autonomous vehicle being in an acceleration state and the longitudinal distance of the second predicted driving trajectory meeting a first longitudinal distance threshold, or in response to the autonomous vehicle being in a deceleration state and the longitudinal distance of the second predicted driving trajectory meeting a second longitudinal distance threshold, it is determined that the second predicted driving trajectory does not conform to the target driving rule, and the longitudinal distance of the second predicted driving trajectory is adjusted based on the acceleration of the autonomous vehicle, wherein the first longitudinal distance threshold is determined based on the current speed of the autonomous vehicle and the acceleration, and the second longitudinal distance threshold is determined based on the acceleration.

14. The method according to any one of claims 1 to 13, wherein, Among the at least one predicted driving trajectory, determining at least one first predicted driving trajectory that conforms to the target driving rule includes: From the at least one predicted driving trajectory, select the number of predicted driving trajectories whose probability values ​​satisfy the probability threshold. For each of the target number of predicted driving trajectories, abnormal trajectory points are removed, or abnormal predicted driving trajectories are removed from the target number of predicted driving trajectories to obtain the at least one first predicted driving trajectory.

15. A method for generating a predicted driving trajectory for an autonomous vehicle, comprising: The user interface displays at least one predicted driving trajectory output by the trajectory prediction model, wherein the trajectory prediction model is used to predict the driving trajectory of the autonomous vehicle. The operation interface displays at least one first predicted driving trajectory that conforms to the target driving rule, wherein the target driving rule is used to guide the autonomous vehicle to drive in a normal driving state on the road segment. In response to a trajectory generation command applied to the operation interface, a second predicted driving trajectory from the at least one first predicted driving trajectory is displayed on the operation interface. The second predicted driving trajectory is used to control the autonomous vehicle to drive to a target lane and is generated based on trajectory points on the first predicted driving trajectory and the target lane. The target lane is the lane that the autonomous vehicle is to drive to and that conforms to the target driving rules. The target lane is determined on the road segment based on the trajectory points on the first predicted driving trajectory. The road coordinate system on the road segment includes a first coordinate axis and a second coordinate axis. Determining the target lane corresponding to the first predicted driving trajectory based on the trajectory points on the first predicted driving trajectory includes: determining the time it takes for the trajectory points on the first predicted driving trajectory to move to each of the candidate target lanes according to the first coordinate axis; determining the offset of the trajectory points on the first predicted driving trajectory when they move to the candidate target lanes according to the second axis according to the time; and determining the target lane based on the offset.

16. A predictive driving trajectory generation device for an autonomous vehicle, comprising: An acquisition unit is used to acquire at least one predicted driving trajectory of an autonomous vehicle; The first determining unit is configured to determine at least one first predicted driving trajectory that conforms to the target driving rule from the at least one predicted driving trajectory, wherein the target driving rule is used to guide the autonomous vehicle to drive in a normal driving state on the road segment; The second determining unit is used to determine the target lane corresponding to the first predicted driving trajectory on the road segment based on the trajectory points on the first predicted driving trajectory, wherein the target lane is the lane that the autonomous vehicle is to drive to and that conforms to the target driving rules; A generation unit is configured to generate a second predicted driving trajectory for the autonomous vehicle based on the trajectory points and the target lane, wherein the second predicted driving trajectory is used to control the autonomous vehicle to drive to the target lane; The road coordinate system on the road segment includes a first coordinate axis and a second coordinate axis. Determining the target lane corresponding to the first predicted driving trajectory based on the trajectory points on the first predicted driving trajectory includes: determining the time it takes for the trajectory points on the first predicted driving trajectory to move to each of the candidate target lanes according to the first coordinate axis; determining the offset of the trajectory points on the first predicted driving trajectory when they move to the candidate target lanes according to the second axis according to the time; and determining the target lane based on the offset.

17. A predictive driving trajectory generation device for an autonomous vehicle, comprising: The first display unit is used to display at least one predicted driving trajectory output by the trajectory prediction model on the operation interface, wherein the trajectory prediction model is used to predict the driving trajectory of the autonomous vehicle. The second display unit is used to display on the operation interface at least one first predicted driving trajectory that conforms to the target driving rule among the at least one predicted driving trajectory, wherein the target driving rule is used to guide the autonomous vehicle to drive in a normal driving state on the road segment; The third display unit is used to respond to a trajectory generation command applied to the operation interface and display a second predicted driving trajectory among the at least one first predicted driving trajectory on the operation interface. The second predicted driving trajectory is used to control the autonomous vehicle to drive to a target lane and is generated based on the trajectory points on the first predicted driving trajectory and the target lane. The target lane is the lane that the autonomous vehicle is to drive to and that conforms to the target driving rules. The target lane is determined on the road segment based on the trajectory points on the first predicted driving trajectory. The road coordinate system on the road segment includes a first coordinate axis and a second coordinate axis. Determining the target lane corresponding to the first predicted driving trajectory based on the trajectory points on the first predicted driving trajectory includes: determining the time it takes for the trajectory points on the first predicted driving trajectory to move to each of the candidate target lanes according to the first coordinate axis; determining the offset of the trajectory points on the first predicted driving trajectory when they move to the candidate target lanes according to the second axis according to the time; and determining the target lane based on the offset.

18. An electronic device comprising: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-15.

19. A non-transitory computer-readable storage medium storing computer instructions, wherein, The computer instructions are used to cause the computer to perform the method according to any one of claims 1-15.

20. A computer program product comprising a computer program that, when executed by a processor, implements the method according to any one of claims 1-15.

21. An autonomous vehicle, including the electronic equipment as claimed in claim 18.

Images