Path planning method, device, medium and vehicle

Abstract

This disclosure provides a path planning method, apparatus, medium, and vehicle, relating to the field of vehicles. The method includes: acquiring trajectory information of a second vehicle ahead of a first vehicle at an intersection; acquiring an initial reference path corresponding to a target type lane in the road where the first vehicle is located, wherein the travel direction indicated by the target type lane is the same as the travel direction of the first vehicle at the intersection, and the initial reference path connects the target type lane and the road opposite the target type lane at the intersection; and adjusting the initial reference path based on the trajectory information of the second vehicle to obtain a target reference path for trajectory planning. This disclosure adjusts the initial reference path by considering the dynamic changes in the traffic flow ahead, using the trajectory of the traffic flow in front of the vehicle. The adjusted target reference path is used for intelligent driving trajectory planning, making the planned intersection travel trajectory safe and reasonable, and improving traffic efficiency.

Description

Technical Field

[0001] This disclosure relates to the field of vehicle technology, and more particularly to a path planning method, apparatus, medium, and vehicle. Background Technology

[0002] In intelligent assisted driving, intersections lack lane markings. Vehicles can plan their trajectories at intersections using virtual lanes generated from map data or sensor data. However, in densely trafficked intersection scenarios, the planned trajectory may lack rationality and safety. Summary of the Invention

[0003] This disclosure provides a route planning method, apparatus, medium, and vehicle to ensure that the generated intersection reference routes are safe and reasonable.

[0004] According to a first aspect of the present disclosure, a path planning method is provided, comprising: Obtain the trajectory information of the second vehicle ahead of the first vehicle at the intersection; Obtain an initial reference path corresponding to the target type lane in the road where the first vehicle is located. The driving direction indicated by the target type lane is the same as the driving direction of the first vehicle at the intersection. The initial reference path is used to connect the target type lane and the road on the opposite side of the target type lane at the intersection. Based on the trajectory information of the second vehicle, the initial reference path is adjusted to obtain a target reference path for trajectory planning.

[0005] Optionally, based on the trajectory information of the second vehicle, the initial reference path is adjusted to obtain a target reference path for trajectory planning, including: The number of targets in a cluster is determined based on the number of lanes of the target type. For each cluster in the target number of clusters, an initial reference path is used as the initial cluster center corresponding to the cluster, and the trajectory information is clustered to obtain the target cluster center corresponding to the cluster. Based on the target cluster center of each cluster, the target reference path corresponding to each cluster is obtained.

[0006] Optionally, using one of the initial reference paths as the initial cluster center for a cluster, clustering the trajectory information to obtain the target cluster center for the cluster includes: The initial cluster center is used as the cluster center to be adjusted for the cluster, and the cluster to which the trajectory information belongs is determined based on the cluster center to be adjusted for the cluster and each position point in the trajectory information. Based on multiple trajectory information in the cluster and the cluster center to be adjusted, a new cluster center corresponding to the cluster is obtained; Based on the cluster centers to be adjusted and the new cluster centers, the target cluster centers corresponding to the cluster are obtained.

[0007] Optionally, obtaining the target cluster center corresponding to the cluster based on the cluster center to be adjusted and the new cluster center includes: When the change value between the cluster center to be adjusted and the new cluster center does not meet the preset stopping condition, the new cluster center is taken as the cluster center to be adjusted, and the step of determining the cluster to which the trajectory information belongs based on the cluster center to be adjusted of the cluster and each position point in the trajectory information is iteratively executed until the change value between the cluster center to be adjusted and the new cluster center meets the preset stopping condition, and the target cluster center corresponding to the cluster is obtained.

[0008] Optionally, determining the cluster to which the trajectory information belongs based on the cluster center to be adjusted and each location point in the trajectory information includes: Projecting multiple location points of the trajectory information onto the reference path corresponding to the cluster center to be adjusted of each cluster, the lateral distance between each location point of the trajectory information and the cluster is obtained; The average lateral distance is obtained based on the lateral distances corresponding to the multiple location points; The cluster to which the trajectory information belongs is determined based on the average lateral distance.

[0009] Optionally, based on multiple trajectory information within the cluster and the cluster centers to be adjusted, a new cluster center corresponding to the cluster is obtained, including: For one of the multiple trajectory information, based on the reference path corresponding to the cluster center to be adjusted and the trajectory information, the lateral offset of the trajectory information at the target position on the reference path is determined, where the target position is any position on the reference path. Based on the lateral offset of the multiple trajectory information at the target position, determine the average lateral offset corresponding to the target position; Based on each target location and the average lateral distance corresponding to the target location, a new cluster center of the cluster is obtained.

[0010] Optionally, the target reference path corresponding to each cluster is obtained based on the target cluster center of each cluster, including: Based on the target cluster center of the cluster, the intermediate reference path corresponding to the cluster is obtained; The intermediate reference path is stabilized and / or smoothed to obtain the target reference path. The stabilization process is used to limit the jitter amplitude of the intermediate reference path, and the smoothing process is used to ensure the continuity of the intermediate reference path.

[0011] Optionally, the intermediate reference path is stabilized to obtain the target reference path, including: Based on the intermediate reference path corresponding to the current time and the intermediate reference path corresponding to the previous time, determine the jump path point on the intermediate reference path corresponding to the previous time and the far path point on the intermediate reference path corresponding to the current time. The far path point is a path point on the intermediate reference path whose distance from the vehicle's location exceeds a preset distance threshold. Interpolation is performed between the transition path point and the remote path point to obtain the target reference path.

[0012] Optionally, before obtaining the trajectory information of the second vehicle ahead of the first vehicle in the intersection, the method further includes: Obtain the orientation angle of the third vehicle in front of the first vehicle relative to the centerline of the lane in which the third vehicle is located, and the lateral distance between the third vehicle and the centerline of the lane in which the first vehicle is located. Based on the navigation information of the first vehicle, the heading angle, and the lateral distance, a third vehicle with the same turning intention as the first vehicle is identified as the second vehicle.

[0013] Optionally, based on the navigation information of the first vehicle, the heading angle, and the lateral distance, determining a third vehicle with the same turning intention as the first vehicle as the second vehicle includes: Based on the navigation information of the first vehicle, determine the turning intention of the first vehicle; Based on the stated steering intention, determine the target orientation angle range and the target lateral distance range; Based on the orientation angle and the target orientation angle range, as well as the lateral distance and the target lateral distance range, a second vehicle among the third vehicles with the same steering intention as the first vehicle is determined.

[0014] Optionally, before adjusting the initial reference path based on the trajectory information of the second vehicle to obtain the target reference path for trajectory planning, the method further includes: Obtain the initial trajectory information of the second vehicle; Based on the trajectory endpoint positions of the initial trajectory information, determine the completeness of the initial trajectory information; When the complete situation indicates that the initial trajectory information is incomplete, the initial trajectory information is completed to obtain the trajectory information.

[0015] According to a second aspect of the present disclosure, a path planning apparatus is provided, comprising: The first acquisition module is configured to acquire trajectory information of a second vehicle ahead of the first vehicle in the intersection. The second acquisition module is configured to acquire an initial reference path corresponding to a target type lane in the road where the first vehicle is located, wherein the driving direction indicated by the target type lane is the same as the driving direction of the first vehicle at the intersection, and the initial reference path is used to connect the target type lane and the road opposite the target type lane at the intersection. The adjustment module is configured to adjust the initial reference path based on the trajectory information of the second vehicle to obtain a target reference path for trajectory planning.

[0016] According to a third aspect of the present disclosure, a computer-readable storage medium is provided having computer program instructions stored thereon, which, when executed by a processor, implement the path planning method described in the first aspect of the present disclosure.

[0017] According to a fourth aspect of the present disclosure, a vehicle is provided, comprising: Storage device for storing computer programs; An execution device is used to execute the computer program to implement the path planning method described in the first aspect of the present disclosure.

[0018] The technical solutions provided by the embodiments of this disclosure may include the following beneficial effects: This disclosure uses the trajectory information of a second vehicle ahead of the first vehicle to adjust the initial reference path corresponding to the target type lane in the road where the first vehicle is located, thus obtaining a target reference path for trajectory planning. The target type lane indicates the same direction of travel as the first vehicle at the intersection, and the initial reference path connects the target type lane to the road on the opposite side of the intersection. In this way, by adjusting the initial reference path based on the trajectory of traffic ahead, the dynamic changes in traffic flow are considered. The adjusted target reference path is used for intelligent driving trajectory planning, ensuring the planned intersection driving trajectory is safe and reasonable, and improving traffic efficiency.

[0019] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0020] The accompanying drawings, which are incorporated in and form a part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure.

[0021] Figure 1 This is a flowchart illustrating a path planning method according to an exemplary embodiment.

[0022] Figure 2 This is a schematic diagram illustrating the stabilization of a reference path according to an exemplary embodiment.

[0023] Figure 3 This is a block diagram illustrating a path planning device according to an exemplary embodiment.

[0024] Figure 4 This is a block diagram illustrating a vehicle according to an exemplary embodiment. Detailed Implementation

[0025] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.

[0026] In the field of intelligent driving, trajectory planning is one of the key technologies to ensure safe and efficient vehicle operation. Path planning methods comprehensively consider the vehicle's dynamic characteristics, road conditions, traffic rules, and the behavior of other traffic participants. In complex intersection scenarios, vehicles are often closely followed by other vehicles when crossing intersections. The virtual lane shape generated from maps or static perception may differ significantly from the actual traffic flow, making it difficult to avoid collisions and increasing the risk of collisions. Furthermore, even when there is no other traffic, an unreasonable virtual lane shape can still result in an unreasonable intersection crossing trajectory, leading to a poor user experience.

[0027] Reference Figure 1 , Figure 1 This is a flowchart illustrating a path planning method according to an exemplary embodiment, such as... Figure 1 As shown, the path planning method includes the following steps.

[0028] In step S101, the trajectory information of the second vehicle in front of the first vehicle in the intersection is obtained.

[0029] In step S102, the initial reference path corresponding to the target type lane in the road where the first vehicle is located is obtained. The driving direction indicated by the target type lane is the same as the driving direction of the first vehicle. The initial reference path is used to connect the target type lane and the road on the opposite side of the target type lane at the intersection.

[0030] In step S103, the initial reference path is adjusted based on the trajectory information of the second vehicle to obtain the target reference path for trajectory planning.

[0031] For example, the first vehicle may be a vehicle with autonomous driving or intelligent driver assistance functions. The second vehicle refers to other vehicles in front of the first vehicle; there may be one or more second vehicles. The second vehicle may be a vehicle in the same lane as the first vehicle, or a vehicle in a lane adjacent to the lane occupied by the first vehicle.

[0032] For example, trajectory information includes path information and time information, with the path information including location and direction information. The trajectory of the second vehicle can be acquired in real time by the perception module of the first vehicle. The perception module may include, but is not limited to, cameras, radar, various sensors, inertial measurement units (IMUs), vehicle communication systems, etc. The first vehicle can directly track the position of the second vehicle at the intersection in real time based on cameras, radar, various sensors, inertial measurement units, etc., to obtain the trajectory information of the second vehicle at the intersection. Alternatively, the first vehicle can request the trajectory information of the second vehicle at the intersection through the vehicle communication system and receive the trajectory information returned by the second vehicle based on the request.

[0033] For example, the road on the same side of the intersection includes multiple lanes, which include multiple types of lanes, such as left-turn lanes, U-turn lanes, straight-ahead lanes, and right-turn lanes. These multiple lanes can be single-type lanes or mixed-type lanes. For instance, the road on the same side has two left-turn lanes, and the leftmost left-turn lane can be both a U-turn lane and a left-turn lane.

[0034] For example, a target type lane is a lane on the road where the first vehicle is currently located, which is in the same direction of travel as the first vehicle at the intersection, and is used to guide the first vehicle and other vehicles through the intersection. The target type lane may include the lane where the first vehicle is located. The target type lane can be a left-turn lane, a U-turn lane, a straight-ahead lane, or a right-turn lane. It is understood that mixed-type lanes on the same side of the road that include the first vehicle's direction of travel at the intersection are also identified as target type lanes. For example, when a vehicle's direction of travel at the intersection is right-turn, and a mixed-type lane indicates both right-turn and straight-ahead directions, that mixed-type road is also identified as a target type road.

[0035] For example, the initial reference path can be obtained from map data, where the map can be a high-precision map. Based on the high-precision map data, virtual lanes at intersections can be generated. The first vehicle can plan its trajectory based on the reference path corresponding to the virtual lane and pass through the intersection based on the trajectory planning result. One end of the virtual lane connects to the lane where the first vehicle is located, and the other end of the virtual lane connects to the opposite road that the first vehicle wants to reach. The initial reference path can be obtained based on the centerline of the virtual lane.

[0036] For example, in a scenario involving a large intersection turn, the initial reference path generated based on a high-precision map might approach a right angle. However, a smaller arc can achieve the same turning objective. Furthermore, the trajectory corresponding to a smaller arc results in a smaller steering angle, which better reflects the driver's actual driving operation and is more reasonable. Therefore, the initial reference path can be adjusted using real-time trajectory information from a second vehicle to obtain a target reference path for actual trajectory planning.

[0037] For example, after obtaining the target reference path, it can be input into the trajectory planner, allowing the planner to plan the trajectory based on the target reference path and obtain the driving trajectory at the intersection. The target reference path generated based on real-time traffic flow can effectively reduce vehicle interaction with other dynamic traffic flows, lowering the risk of collisions. Furthermore, the target reference line generated based on real-time traffic flow can mimic the detour behavior of the vehicle in front. In case of abnormal situations at the intersection, the vehicle can follow the trajectory of the vehicle in front to complete the detour, improving traffic efficiency.

[0038] This disclosure uses the trajectory information of a second vehicle ahead of the first vehicle to adjust the initial reference path corresponding to the target type lane in the road where the first vehicle is located, thus obtaining a target reference path for trajectory planning. The target type lane indicates the same direction of travel as the first vehicle at the intersection, and the initial reference path connects the target type lane to the road on the opposite side of the intersection. In this way, by adjusting the initial reference path based on the trajectory of traffic ahead, the dynamic changes in traffic flow are considered. The adjusted target reference path is used for intelligent driving trajectory planning, ensuring the planned intersection driving trajectory is safe and reasonable, and improving traffic efficiency.

[0039] As an optional embodiment, the initial reference path is adjusted based on the trajectory information of the second vehicle to obtain a target reference path for trajectory planning, including: The number of targets in a cluster is determined based on the number of lanes of the target type. For each cluster in the target number of clusters, an initial reference path is used as the initial cluster center corresponding to the cluster, and the trajectory information is clustered to obtain the target cluster center corresponding to the cluster. Based on the target cluster center of each cluster, the target reference path corresponding to each cluster is obtained.

[0040] For example, clustering is a statistical method that groups data points into groups such that data points within the same group are highly similar to each other, while the similarity between different groups is low. Clustering algorithms can include k-means clustering, Gaussian Mixture Models, Fuzzy C-Means Clustering, and some neural network-based clustering algorithms can also be used to cluster trajectory information.

[0041] For example, a cluster is a grouping of trajectories, with each cluster containing information about one or more trajectories. A cluster center is the center or average location of a cluster, used to represent that cluster. The cluster center of each cluster can be the average location of all trajectory information, used to represent the characteristics of that cluster. The target cluster center is the final cluster center of each cluster after the clustering process.

[0042] For example, the target number of clusters can be determined based on the number of lanes of the target type. The target number can correspond to the number of lanes of the target type; for example, the target number of clusters can be the same as the number of lanes of the target type. Alternatively, the target number of clusters can be based on the number of lanes of the target type by adding other clusters. For example, in addition to the clusters corresponding to the lanes of the target type, an extra cluster can be added to cluster other vehicles whose trajectories change abruptly in the middle of the intersection; this is not limited here.

[0043] For example, the initial cluster center of each cluster is the initial reference path of the corresponding target type lane. After clustering the trajectory information based on the initial cluster center of the cluster, the target cluster center obtained is also a reference path. The reference path corresponding to the final target cluster center can be directly used as the target reference path.

[0044] For example, based on the initial reference path corresponding to the target type lane, a preset number of cluster centers are determined, with each cluster center corresponding to a cluster. Specifically, the cluster to which each trajectory information belongs can be determined based on the average deviation between each trajectory information and each cluster center. For each cluster, a new cluster center can be calculated based on the trajectory information within the cluster, representing the average position of all points in the cluster.

[0045] For example, based on the number of lanes of the target type, the number of clusters is determined, and after initializing the cluster center of each cluster, the trajectory information is clustered, assigning each location point in the trajectory information to the nearest cluster center to form different clusters. For each cluster, new cluster centers for all trajectory information assigned to that cluster can be determined, and based on the target cluster centers of each cluster, the target reference path corresponding to each cluster is generated.

[0046] Based on this, the initial reference path is adjusted using the actual trajectory information of the second vehicle to obtain an optimized target reference path. This target reference path better reflects actual traffic flow, helps avoid potential collisions and conflicts, improves driving safety, and enhances the rationality of the target reference path's shape. Furthermore, the decision-making process using cluster analysis can more comprehensively consider the trajectory information of multiple preceding vehicles, improving the accuracy and robustness of trajectory planning.

[0047] As an optional embodiment, the initial reference path is used as the initial cluster center of a cluster, and the trajectory information is clustered to obtain the target cluster center corresponding to the cluster, including: The initial cluster center is used as the cluster center to be adjusted for the cluster, and the cluster to which the trajectory information belongs is determined based on the cluster center to be adjusted for the cluster and each position point in the trajectory information. Based on multiple trajectory information in the cluster, a new cluster center corresponding to the cluster is obtained; Based on the cluster centers to be adjusted and the new cluster centers, the target cluster centers corresponding to the cluster are obtained.

[0048] For example, the cluster centers to be adjusted are those that need to be updated during the clustering process. In the first iteration, the initial cluster centers, i.e., the initial reference paths, can be used as the cluster centers to be adjusted. The new cluster centers are the new cluster centers of the clusters calculated after the clustering algorithm iterations, and can represent the updated reference paths.

[0049] For example, using the initial reference path as the initial cluster center, each location point in the trajectory information can be assigned to a different cluster based on the initial cluster center. For the trajectory information assigned to each cluster, a new cluster center can be calculated based on the trajectory information in that cluster.

[0050] For example, the target cluster center corresponding to the cluster can be obtained based on the new cluster center. For instance, during the first iteration, if the iteration stopping condition is met based on the cluster center to be adjusted and the new cluster center, the new cluster center can be used as the target cluster center, and the target reference path can be obtained based on the reference path corresponding to the target cluster center.

[0051] For example, the cluster to which the trajectory information belongs can be determined based on the distance between the cluster center to be adjusted and each location point in the trajectory information. For a trajectory information, the lateral distance between each location point and the reference path corresponding to the cluster center to be adjusted can be determined based on multiple location points corresponding to the trajectory information, and the average of the lateral distances corresponding to multiple location points can be calculated as the distance information between the trajectory information and the cluster center to be adjusted; based on the distance information between the cluster center to be adjusted and the trajectory information, the cluster to which the trajectory information belongs can be determined.

[0052] Based on this, cluster analysis can be used to cluster the trajectory information of the second vehicle ahead of the first vehicle, update the initial reference path, and obtain a target reference path for trajectory planning, which can improve the accuracy of trajectory planning. The path optimized by cluster analysis may follow the vehicle in front to change lanes and make adjustments, and can also reduce unnecessary detours and improve traffic efficiency.

[0053] As an optional embodiment, obtaining the target cluster center corresponding to the cluster based on the cluster center to be adjusted and the new cluster center includes: When the change value between the cluster center to be adjusted and the new cluster center does not meet the preset stopping condition, the new cluster center is taken as the cluster center to be adjusted, and the step of determining the cluster to which the trajectory information belongs based on the cluster center to be adjusted of the cluster and each position point in the trajectory information is iteratively executed until the change value between the cluster center to be adjusted and the new cluster center meets the preset stopping condition, and the target cluster center corresponding to the cluster is obtained.

[0054] For example, the change value is the difference between cluster centers after two adjacent iterations, which can be used to measure the convergence of the clustering process. Specifically, the preset stopping condition can be determined to be met when the change value between the cluster center to be adjusted and the new cluster center is less than a preset change threshold; or, the preset stopping condition can be determined to be met when the cluster centers no longer change, i.e., when the change value between the cluster center to be adjusted and the new cluster center is 0.

[0055] For example, if the change value between the cluster center to be adjusted and the new cluster center does not meet the preset stopping condition, it indicates that the new cluster center does not meet the convergence requirement. The steps of determining the clusters for trajectory information and updating the cluster centers can be performed iteratively. After each iteration, the change value between the cluster centers after two adjacent iterations is calculated, i.e., the change value between the cluster centers before the current iteration and the new cluster centers after the current iteration, to determine whether the clustering is convergent. This process continues until the change value meets the preset iteration stopping condition, and the finally determined cluster centers are used to generate an intermediate reference path.

[0056] It is understandable that, since the difference between the cluster centers to be adjusted before this iteration and the new cluster centers after this iteration is small, both the cluster centers to be adjusted and the new cluster centers can be used as the target cluster centers. Alternatively, the mean of the cluster centers to be adjusted and the new cluster centers can be used as the target cluster centers; there is no restriction here.

[0057] Based on this, the target cluster center can be determined more accurately through the iterative process of cluster center determination, which helps to avoid potential collisions and conflicts, improve driving safety, and enable the target reference path determined based on the target cluster center to adapt to the real-time changing trajectory information of the vehicle in front of the first vehicle, making path planning more flexible and dynamic. The iterative process reduces the dependence of the clustering results on the initial conditions, thus improving the robustness of the algorithm.

[0058] As an optional embodiment, determining the cluster to which the trajectory information belongs based on the cluster center to be adjusted and each location point in the trajectory information includes: Projecting multiple location points of the trajectory information onto the reference path corresponding to the cluster center to be adjusted of each cluster, the lateral distance between each location point of the trajectory information and the cluster is obtained; The average lateral distance is obtained based on the lateral distances corresponding to the multiple location points; The cluster to which the trajectory information belongs is determined based on the average lateral distance.

[0059] For example, lateral distance is the perpendicular distance from a location point in the trajectory information to a reference path. Average lateral distance is the average of the perpendicular distances from multiple location points in the trajectory information to the reference path. Projection can be the process of mapping location points on the trajectory information to a cluster center or a reference path. For example, projecting location points in the trajectory information onto the reference path corresponding to a cluster center allows for the calculation of the distance between the location point and the reference path.

[0060] For example, projection can be performed through isometric mapping, Frenet frame transformation, etc. The Frenet frame transformation, also known as the Frenet-Serret coordinate system, can be used to describe the local coordinate system of a point on a curve. The three axes of the Frenet-Serret coordinate system are dynamically changing, varying with the curve. For instance, a Frenet frame can be used to represent a vehicle's position (s) and lateral offset (l) on the road, where s is the arc length along the road and l is the lateral distance from the vehicle to the road centerline.

[0061] For example, by projecting multiple location points in the trajectory information onto the reference path corresponding to the cluster center of each cluster, the lateral distance between each location point and the reference path corresponding to the cluster center to be adjusted in a cluster can be obtained. The average lateral distance is then calculated by averaging the lateral distances corresponding to multiple location points. Based on the average lateral distance between the cluster center to be adjusted in each cluster and the trajectory information, the cluster to which the trajectory information belongs can be determined.

[0062] Therefore, by using the average lateral distance as a distance metric, even if the observed trajectory lengths of the second vehicle differ due to the blind spot of the first vehicle, the impact on the trajectory clustering results is minimal. In other words, using the average lateral distance as a distance metric provides robustness against noise and outliers, resulting in more stable clustering results. Consequently, trajectory information points can be more accurately assigned to the best-matching clusters, improving clustering accuracy.

[0063] As an optional embodiment, based on multiple trajectory information within the cluster and the cluster centers to be adjusted, a new cluster center corresponding to the cluster is obtained, including: For one of the multiple trajectory information, based on the reference path corresponding to the cluster center to be adjusted and the trajectory information, the lateral offset of the trajectory information at the target position on the reference path is determined, where the target position is any position on the reference path. Based on the lateral offset of the multiple trajectory information at the target position, determine the average lateral offset corresponding to the target position; Based on each target location and the average lateral distance corresponding to the target location, a new cluster center of the cluster is obtained.

[0064] For example, the target location is any location on the reference path. Lateral offset is the lateral distance of a location point in the trajectory information relative to a location point on the reference path. Specifically, by projecting the trajectory information onto the reference path, the location point in the trajectory information corresponding to the target location can be determined, and the lateral offset is determined based on the target location and the location point in the trajectory information corresponding to the target location. Based on the lateral offsets of multiple trajectory information points corresponding to the target location, the average lateral offset of the target location can be determined.

[0065] For example, a new cluster center of the cluster can be obtained by performing coordinate transformation based on each of the target locations and the average lateral distance corresponding to the target locations. Coordinate transformation is the process of converting the location points in the trajectory information from one coordinate system to another based on the average lateral offset, such as from the Frenet-Serret coordinate system to the Cartesian coordinate system.

[0066] For example, for each trajectory information in a cluster, its lateral offset at the target position on the reference path is calculated. For each target position on the reference path, the average lateral offset of all trajectory information at that position can be calculated, and the cluster center is updated based on each target position and its corresponding average lateral offset to make it closer to the average position of the trajectory information.

[0067] Based on this, by updating the target position of the reference path based on the average lateral offset of multiple trajectory points relative to the target position of the reference path, a new cluster center can be obtained, which can improve the accuracy and robustness of clustering.

[0068] As an optional embodiment, obtaining the target reference path corresponding to each cluster based on the target cluster center of each cluster includes: Based on the target cluster center of the cluster, the intermediate reference path corresponding to the cluster is obtained; The intermediate reference path is stabilized and / or smoothed to obtain the target reference path. The stabilization process is used to limit the jitter amplitude of the intermediate reference path, and the smoothing process is used to ensure the continuity of the intermediate reference path.

[0069] For example, the intermediate reference path is the reference path corresponding to the target cluster center obtained after cluster analysis of the trajectory information of the second vehicle. The intermediate reference path is generally a coarse reference line and may exhibit significant jitter or sharp turning points. Stabilization processing can involve polynomial or asymptotic interpolation between two intermediate reference paths at adjacent time points. It can also utilize neural network models for predictive control of path points, predicting future trajectories and optimizing control inputs to reduce abrupt changes. Smoothing processing can employ smoothing algorithms such as polynomial fitting, spline interpolation, Bézier curves, and Kalman filtering. The specific methods used for stabilization and smoothing processing are not limited here.

[0070] Additionally, smoothing strategies can be applied at each turning point or key point of the trajectory to ensure the stability and smoothness of each segment. Alternatively, optimization algorithms such as genetic algorithms and particle swarm optimization can be used to optimize intermediate reference paths while simultaneously ensuring the smoothness and stability of the target reference path.

[0071] For example, stabilization processing reduces path jitter and abrupt changes, improving path stability and reliability. Smoothing processing reduces sharp turns and lane changes, enhancing driving comfort and safety. The intermediate reference path, after stabilization and smoothing, can be used as a target reference path for trajectory planning in autonomous vehicles.

[0072] Based on this, stabilization and smoothing processes make the system more robust to sensor noise and data anomalies, reducing path jitter and abrupt changes for a more comfortable ride. This also reduces the risk of vehicle accidents and improves driving safety. Furthermore, paths based on smoothing and stabilization are easier for vehicles to track, reducing the need for high-precision control.

[0073] As an optional embodiment, stabilizing the intermediate reference path to obtain the target reference path includes: Based on the intermediate reference path corresponding to the current time and the intermediate reference path corresponding to the previous time, determine the jump path point on the intermediate reference path corresponding to the previous time and the far path point on the intermediate reference path corresponding to the current time. The far path point is a path point on the intermediate reference path whose distance from the vehicle's location exceeds a preset distance threshold. Interpolation is performed between the transition path point and the remote path point to obtain the target reference path.

[0074] For example, a path jump point is a point where the path changes abruptly across consecutive moments. This can be caused by sensor errors, environmental changes, or data noise. For instance, the trajectory information of the second vehicle observed by the first vehicle may be assigned to different clusters at different times, causing the calculated cluster centers to jump. Another example is the movement of the second vehicle itself, which causes changes in the shape of the traffic flow.

[0075] For example, a remote path point is a path point on an intermediate reference path whose distance from the vehicle's current position exceeds a preset threshold. The preset distance threshold is used to determine the distance limit for remote path points; path points exceeding this distance are identified as remote path points.

[0076] For example, by comparing the intermediate reference path at the current moment with the intermediate reference path at the previous moment, the abrupt path points in the intermediate reference path corresponding to the previous moment can be identified. Then, based on the vehicle's current position, the far-end path point on the intermediate reference path corresponding to the current moment can be determined. Interpolating between the identified abrupt path points and the far-end path points can generate a smoother and more stable path.

[0077] Reference Figure 3 This diagram illustrates an interpolation method. Here, time t0 represents the previous time, and time t1 represents the current time. Correspondingly, point A represents the far-end path point on the intermediate reference path corresponding to the current time; point B represents the current position of the first vehicle; and point C represents the transition path point on the intermediate reference path corresponding to the previous time. For example... Figure 3As shown, if no stabilization processing is performed on the intermediate reference path, when switching from the intermediate reference path corresponding to t0 to the intermediate reference path corresponding to t1, the vehicle will directly travel from point C to the intermediate reference path corresponding to t1, potentially requiring a large steering angle for the switch. Therefore, interpolation can be performed between points A and C to obtain... Figure 3 The dashed path shown is used to obtain a more stable target reference path.

[0078] Based on this, by identifying and stabilizing waypoints, abrupt changes in the path are reduced, path stability is improved, and passenger comfort is enhanced.

[0079] As an optional embodiment, before obtaining the trajectory information of the second vehicle ahead of the first vehicle at the intersection, the method further includes: Obtain the orientation angle of the third vehicle in front of the first vehicle relative to the centerline of the lane in which the third vehicle is located, and the lateral distance between the third vehicle and the centerline of the lane in which the first vehicle is located. Based on the navigation information of the first vehicle, the heading angle, and the lateral distance, a third vehicle with the same turning intention as the first vehicle is identified as the second vehicle.

[0080] For example, a third vehicle is a vehicle in front of the first vehicle, and there can be more than one third vehicle. The third vehicle can be in the same lane as the first vehicle or in a different lane. The heading angle of the third vehicle is the angle of the third vehicle relative to the center line of its lane. The lateral distance is the vertical distance between the center lines of the lanes occupied by the third vehicle and the first vehicle, and can characterize the lateral distance between the three vehicles. The heading angle and lateral distance information of the third vehicle in front of the first vehicle can be obtained through the first vehicle's onboard sensors such as radar, lidar, or cameras. Navigation information is data provided by the first vehicle's navigation system, including position, speed, and direction of travel.

[0081] For example, based on the navigation information of the first vehicle, the turning intention of the first vehicle at the intersection ahead can be determined. Based on the heading angle of the third vehicle and the lateral distance between the third vehicle and the centerline of the lane where the first vehicle is located, the turning intention of the third vehicle at the intersection ahead can be determined. The turning intention may include a straight-ahead intention, a left-turn intention, a U-turn intention, and a right-turn intention, etc.

[0082] Based on this, by combining navigation information and sensor data, the location and driving intention of the third vehicle can be identified more accurately. This allows the third vehicle with the same turning intention as the first vehicle to be identified as the second vehicle to be clustered, reducing redundant data when planning the trajectory of the first vehicle and improving the accuracy and reliability of the clustering analysis.

[0083] As an optional embodiment, determining a third vehicle with the same turning intention as the first vehicle as the second vehicle based on the navigation information of the first vehicle, the heading angle, and the lateral distance includes: Based on the navigation information of the first vehicle, determine the turning intention of the first vehicle; Based on the stated steering intention, determine the target orientation angle range and the target lateral distance range; Based on the orientation angle and the target orientation angle range, as well as the lateral distance and the target lateral distance range, a second vehicle among the third vehicles with the same steering intention as the first vehicle is determined.

[0084] For example, the target steering angle range is the expected range of steering angles that match the steering intention of the first vehicle. The target lateral distance range is the expected range of lateral distances that match the steering intention of the first vehicle. Both the target steering angle range and the target lateral distance range can be preset according to different steering intentions.

[0085] For example, taking a lane centerline of 0° and 0m as an example, when the first vehicle's turning intention is to turn right, the target orientation angle can be determined to be between 0 and 45°, and the target lateral distance between 0 and 5m. For example, when the first vehicle's turning intention is to turn left or make a U-turn, the target orientation angle can be determined to be between 135° and 0°, and the target lateral distance between -5 and -0m. For example, when the first vehicle's turning intention is to go straight, the target orientation angle can be determined to be between 135° and 45°, and the target lateral distance between -5m and 5m. These are merely examples and are not intended to limit this embodiment.

[0086] For example, based on the navigation information of the first vehicle, such as map data, current lane information, and a preset driving route, the steering intention of the first vehicle can be determined. Based on the steering intention of the first vehicle, the corresponding azimuth angle range and lateral distance range are determined. Then, based on the azimuth angle and lateral distance of the third vehicle, and the target azimuth angle range and target lateral distance range of the first vehicle, a second vehicle with the same steering intention as the first vehicle is identified. Specifically, if the azimuth angle of the third vehicle is within the target azimuth angle range and / or the lateral distance is within the target lateral distance range, the third vehicle is determined to be the second vehicle with the same steering intention as the first vehicle.

[0087] As an optional embodiment, before adjusting the initial reference path based on the trajectory information of the second vehicle to obtain the target reference path for trajectory planning, the method further includes: Obtain the initial trajectory information of the second vehicle; Based on the trajectory endpoint positions of the initial trajectory information, determine the completeness of the initial trajectory information; When the complete situation indicates that the initial trajectory information is incomplete, the initial trajectory information is completed to obtain the trajectory information.

[0088] For example, the initial trajectory information is based on trajectory data of the second vehicle directly obtained from a camera or radar. Since the second vehicle's trajectory information may be partially within the first vehicle's blind spot, or the second vehicle may not have completely passed the intersection, the second vehicle's trajectory information may be incomplete. The trajectory endpoint positions are the start and end positions of the second vehicle recorded in the initial trajectory information. Completeness, in this case, indicates whether the initial trajectory information is complete.

[0089] For example, the completeness of the initial trajectory information can be determined based on the trajectory endpoint locations. For instance, it can be determined whether the trajectory endpoint location is connected to a target type lane, and whether the trajectory endpoint location is connected to the road opposite the target type lane at the intersection. If at least one condition is not met, the initial trajectory information is determined to be incomplete.

[0090] For example, the initial trajectory information can be completed to obtain the trajectory information. Completing the initial trajectory information can be done using methods such as linear interpolation, polynomial fitting, or Bézier curves. For instance, a second-order Bézier curve can be used to complete the initial trajectory information.

[0091] Therefore, by completing the incomplete trajectory information, the integrity of the data is ensured, providing more accurate data support for subsequent path planning. The method of completing trajectory information improves the algorithm's robustness to data gaps and anomalies.

[0092] As a specific embodiment, the path planning method may include the following steps: S1: Identify a vehicle ahead with the same steering intention as the vehicle ahead. Based on the target heading angle range and target lateral distance range corresponding to the vehicle ahead's steering intention, the vehicle ahead can satisfy at least one of the following: a) The difference between the heading angle of the vehicle ahead and the heading angle corresponding to the center line of the lane where the vehicle ahead is located is within the target heading angle range; b) The lateral distance between the vehicle ahead and the center line of the lane where the vehicle ahead is located is within the target lateral distance range.

[0093] S2: Record the trajectory information of the preceding vehicle determined in step S1, and perform trajectory clustering. Trajectory clustering can be performed using k-means, and the main steps are as follows: S21: Determine the k value for k-means clustering. k can be determined based on the number of lanes on the same side of the intersection that are traveling in the same direction as the vehicle, or the k value can be appropriately increased based on other types of traffic flow corresponding to the vehicle changing lanes in front of it in the intersection.

[0094] S22: Initialization, the lane centerlines corresponding to the k lanes will be used as the initial values ​​of the cluster centers of each of the k clusters, where the lane centerlines are the initial reference paths.

[0095] S23: Update the cluster center value for each cluster. First, based on the trajectory information and cluster centers, cluster the trajectory information to determine the cluster to which each trajectory belongs. Then, project each trajectory in a cluster onto the reference path corresponding to that cluster, i.e., perform a Frenet frame transformation. Then, the lateral offset (l) of each position (s) of the lane centerline relative to the reference path can be obtained. Then, a series of (s, l) coordinate points can be transformed to Cartesian coordinates using the Frenet frame to obtain a coarse reference line in Cartesian coordinates. Updating the mean completes one iteration. If the change in the cluster center between the current iteration and the previous iteration is less than a set threshold, the trajectory clustering converges.

[0096] In some cases, due to blind spots or obstacles not completely passing through the intersection, the observed trajectory information of the vehicle in front is inconsistent in length. For the unobserved part, before clustering the trajectory information based on the trajectory information and cluster centers, a second-order Bézier curve can be used to complete the two endpoints of the observed trajectory information with the intersection entrance and exit positions.

[0097] S3: Post-processing. The cluster centers of the clusters after trajectory clustering convergence are coarse reference lines, which can be used for stabilization and smoothing, and output as reference paths that can be used by the planner.

[0098] In trajectory clustering, unstable factors may include, for example, the trajectory information of the second vehicle observed by the first vehicle may be assigned to different clusters at different times, causing jumps in the calculated cluster centers. Another example is the swaying of the second vehicle itself, causing changes in the traffic flow shape. Figure 3 As shown, to reduce the impact of path jitter, a strategy of fewer updates at the near end and more updates at the far end is adopted. Finally, the reference path can be smoothed based on the conventional reference line smoothing method (FemPosDeviationSmoother) to obtain a reference path that can be used by the planner.

[0099] This disclosure utilizes reference paths generated from real-time traffic flow information, naturally reducing interaction with other dynamic traffic flows and lowering the risk of collisions. Furthermore, the reference lines generated based on real-time traffic flow information can mimic the detour behavior of the vehicle in front. In the event of abnormal situations at intersections, the vehicle can follow the trajectory of the vehicle in front to complete a detour, improving traffic efficiency.

[0100] Reference Figure 3 , Figure 3 This is a block diagram illustrating a path planning device 300 according to an exemplary embodiment. Figure 3 As shown, the path planning device 300 includes a first acquisition module 301, a second acquisition module 302, an adjustment module 303, and a planning module 304.

[0101] The first acquisition module 301 is configured to acquire trajectory information of a second vehicle ahead of the first vehicle at the intersection. The second acquisition module 302 is configured to acquire an initial reference path corresponding to a target type lane in the road where the first vehicle is located, wherein the driving direction indicated by the target type lane is the same as the driving direction of the first vehicle at the intersection, and the initial reference path is used to connect the target type lane and the road opposite the target type lane at the intersection. The adjustment module 303 is configured to adjust the initial reference path based on the trajectory information of the second vehicle to obtain a target reference path for trajectory planning.

[0102] As an optional embodiment, the adjustment module 303 is configured to: The number of targets in a cluster is determined based on the number of lanes of the target type. For each cluster in the target number of clusters, an initial reference path is used as the initial cluster center corresponding to the cluster, and the trajectory information is clustered to obtain the target cluster center corresponding to the cluster. Based on the target cluster center of each cluster, the target reference path corresponding to each cluster is obtained.

[0103] As an optional embodiment, the adjustment module 303 is configured to: The initial cluster center is used as the cluster center to be adjusted for the cluster, and the cluster to which the trajectory information belongs is determined based on the cluster center to be adjusted for the cluster and each position point in the trajectory information. Based on multiple trajectory information in the cluster and the cluster center to be adjusted, a new cluster center corresponding to the cluster is obtained; Based on the cluster centers to be adjusted and the new cluster centers, the target cluster centers corresponding to the cluster are obtained.

[0104] As an optional embodiment, the adjustment module 303 is configured to: When the change value between the cluster center to be adjusted and the new cluster center does not meet the preset stopping condition, the new cluster center is taken as the cluster center to be adjusted, and the step of determining the cluster to which the trajectory information belongs based on the cluster center to be adjusted of the cluster and each position point in the trajectory information is iteratively executed until the change value between the cluster center to be adjusted and the new cluster center meets the preset stopping condition, and the target cluster center corresponding to the cluster is obtained.

[0105] As an optional embodiment, the adjustment module 303 is configured to: Projecting multiple location points of the trajectory information onto the reference path corresponding to the cluster center to be adjusted of each cluster, the lateral distance between each location point of the trajectory information and the cluster is obtained; The average lateral distance is obtained based on the lateral distances corresponding to the multiple location points; The cluster to which the trajectory information belongs is determined based on the average lateral distance.

[0106] As an optional embodiment, the adjustment module 303 is configured to: For one of the multiple trajectory information, based on the reference path corresponding to the cluster center to be adjusted and the trajectory information, the lateral offset of the trajectory information at the target position on the reference path is determined, where the target position is any position on the reference path. Based on the lateral offset of the multiple trajectory information at the target position, determine the average lateral offset corresponding to the target position; Based on each target location and the average lateral distance corresponding to the target location, a new cluster center of the cluster is obtained.

[0107] As an optional embodiment, the adjustment module 303 is configured to: Based on the target cluster center of the cluster, the intermediate reference path corresponding to the cluster is obtained; The intermediate reference path is stabilized and / or smoothed to obtain the target reference path. The stabilization process is used to limit the jitter amplitude of the intermediate reference path, and the smoothing process is used to ensure the continuity of the intermediate reference path.

[0108] As an optional embodiment, the adjustment module 303 is configured to: Based on the intermediate reference path corresponding to the current time and the intermediate reference path corresponding to the previous time, determine the jump path point on the intermediate reference path corresponding to the previous time and the far path point on the intermediate reference path corresponding to the current time. The far path point is a path point on the intermediate reference path whose distance from the vehicle's location exceeds a preset distance threshold. Interpolation is performed between the transition path point and the remote path point to obtain the target reference path.

[0109] As an optional embodiment, the path planning device 300 is also configured to: Obtain the orientation angle of the third vehicle in front of the first vehicle relative to the centerline of the lane in which the third vehicle is located, and the lateral distance between the third vehicle and the centerline of the lane in which the first vehicle is located. Based on the navigation information of the first vehicle, the heading angle, and the lateral distance, a third vehicle with the same turning intention as the first vehicle is identified as the second vehicle.

[0110] As an optional embodiment, the path planning device 300 is also configured to: Based on the navigation information of the first vehicle, determine the turning intention of the first vehicle; Based on the stated steering intention, determine the target orientation angle range and the target lateral distance range; Based on the orientation angle and the target orientation angle range, as well as the lateral distance and the target lateral distance range, a second vehicle among the third vehicles with the same steering intention as the first vehicle is determined.

[0111] As an optional embodiment, the path planning device 300 is also configured to: Obtain the initial trajectory information of the second vehicle; Based on the trajectory endpoint positions of the initial trajectory information, determine the completeness of the initial trajectory information; When the complete situation indicates that the initial trajectory information is incomplete, the initial trajectory information is completed to obtain the trajectory information.

[0112] Regarding the path planning device 300 in the above embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments related to the path planning method, and will not be elaborated here.

[0113] Based on the same inventive concept, this disclosure also provides a computer-readable storage medium having computer program instructions stored thereon, which, when executed by a processor, implement the path planning method provided in this disclosure.

[0114] Based on the same inventive concept, this disclosure also provides a vehicle, comprising: Storage device for storing computer programs; An execution device is used to execute the computer program to implement the path planning method provided in this disclosure.

[0115] In another exemplary embodiment, a computer program product is also provided, the computer program product comprising a computer program executable by a programmable device, the computer program having a code portion for performing the path planning method described above when executed by the programmable device.

[0116] Figure 4 This is a block diagram illustrating a vehicle 700 according to an exemplary embodiment. For example, vehicle 700 can be a hybrid vehicle, a non-hybrid vehicle, an electric vehicle, a fuel cell vehicle, or other types of vehicle. Vehicle 700 can be an autonomous vehicle or a semi-autonomous vehicle.

[0117] Reference Figure 4 The vehicle 700 may include various subsystems, such as an infotainment system 710, a perception system 720, a decision control system 730, a drive system 740, and a computing platform 750. The vehicle 700 may also include more or fewer subsystems, and each subsystem may include multiple components. Furthermore, each subsystem and each component of the vehicle 700 can be interconnected via wired or wireless means.

[0118] In some embodiments, the infotainment system 710 may include a communication system, an entertainment system, and a navigation system, etc.

[0119] The perception system 720 may include several sensors for sensing information about the environment surrounding the vehicle 700. For example, the perception system 720 may include a global positioning system (which may be GPS, BeiDou, or other positioning systems), an inertial measurement unit (IMU), lidar, millimeter-wave radar, ultrasonic radar, and a camera device.

[0120] The decision control system 730 may include a computing system, a vehicle controller, a steering system, a throttle, and a braking system.

[0121] The drive system 740 may include components that provide powered motion to the vehicle 700. In one embodiment, the drive system 740 may include an engine, an energy source, a transmission system, and wheels. The engine may be one or a combination of internal combustion engines, electric motors, and compressed air engines. The engine is capable of converting energy provided by the energy source into mechanical energy.

[0122] Some or all of the functions of vehicle 700 are controlled by computing platform 750. Computing platform 750 may include at least one processor 751 and memory 752, and processor 751 may execute instructions 753 stored in memory 752.

[0123] Processor 751 can be any conventional processor, such as a commercially available CPU. Processors may also include graphics processing units (GPUs), field-programmable gate arrays (FPGAs), systems-on-chips (SoCs), application-specific integrated circuits (ASICs), or combinations thereof.

[0124] The memory 752 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0125] In addition to instruction 753, memory 752 can also store data, such as road maps, route information, vehicle position, direction, speed, and other data. The data stored in memory 752 can be used by computing platform 750.

[0126] In this embodiment of the disclosure, processor 751 may execute instruction 753 to complete all or part of the steps of the path planning method described above.

[0127] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the specification and practice of this disclosure. This disclosure is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the following claims.

[0128] It should be understood that this disclosure is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this disclosure is limited only by the appended claims.

Claims

1. A path planning method, characterized in that, include: Obtain the trajectory information of the second vehicle ahead of the first vehicle at the intersection; Obtain an initial reference path corresponding to the target type lane in the road where the first vehicle is located. The driving direction indicated by the target type lane is the same as the driving direction of the first vehicle at the intersection. The initial reference path is used to connect the target type lane and the road on the opposite side of the target type lane at the intersection. Based on the trajectory information of the second vehicle, the initial reference path is adjusted to obtain a target reference path for trajectory planning.

2. The method according to claim 1, characterized in that, Based on the trajectory information of the second vehicle, the initial reference path is adjusted to obtain a target reference path for trajectory planning, including: The number of targets in a cluster is determined based on the number of lanes of the target type. For each cluster in the target number of clusters, an initial reference path is used as the initial cluster center corresponding to the cluster, and the trajectory information is clustered to obtain the target cluster center corresponding to the cluster. Based on the target cluster center of each cluster, the target reference path corresponding to each cluster is obtained.

3. The method according to claim 2, characterized in that, Using the initial reference path as the initial cluster center for a cluster, the trajectory information is clustered to obtain the target cluster center for the cluster, including: The initial cluster center is used as the cluster center to be adjusted for the cluster, and the cluster to which the trajectory information belongs is determined based on the cluster center to be adjusted for the cluster and each position point in the trajectory information. Based on multiple trajectory information in the cluster and the cluster center to be adjusted, a new cluster center corresponding to the cluster is obtained; Based on the cluster centers to be adjusted and the new cluster centers, the target cluster centers corresponding to the cluster are obtained.

4. The method according to claim 3, characterized in that, Based on the cluster centers to be adjusted and the new cluster centers, the target cluster centers corresponding to the cluster are obtained, including: When the change value between the cluster center to be adjusted and the new cluster center does not meet the preset stopping condition, the new cluster center is taken as the cluster center to be adjusted, and the step of determining the cluster to which the trajectory information belongs based on the cluster center to be adjusted of the cluster and each position point in the trajectory information is iteratively executed until the change value between the cluster center to be adjusted and the new cluster center meets the preset stopping condition, and the target cluster center corresponding to the cluster is obtained.

5. The method according to claim 3, characterized in that, Based on the cluster centers to be adjusted for the given cluster and each location point in the trajectory information, determine the cluster to which the trajectory information belongs, including: Projecting multiple location points of the trajectory information onto the reference path corresponding to the cluster center to be adjusted of each cluster, the lateral distance between each location point of the trajectory information and the cluster is obtained; The average lateral distance is obtained based on the lateral distances corresponding to the multiple location points; The cluster to which the trajectory information belongs is determined based on the average lateral distance.

6. The method according to claim 3, characterized in that, Based on multiple trajectory information within the cluster and the cluster centers to be adjusted, new cluster centers corresponding to the cluster are obtained, including: For one of the multiple trajectory information, based on the reference path corresponding to the cluster center to be adjusted and the trajectory information, the lateral offset of the trajectory information at the target position on the reference path is determined, where the target position is any position on the reference path. Based on the lateral offset of the multiple trajectory information at the target position, determine the average lateral offset corresponding to the target position; Based on each target location and the average lateral distance corresponding to the target location, a new cluster center of the cluster is obtained.

7. The method according to claim 2, characterized in that, Based on the target cluster center of each cluster, the target reference path corresponding to each cluster is obtained, including: Based on the target cluster center of the cluster, the intermediate reference path corresponding to the cluster is obtained; The intermediate reference path is stabilized and / or smoothed to obtain the target reference path. The stabilization process is used to limit the jitter amplitude of the intermediate reference path, and the smoothing process is used to ensure the continuity of the intermediate reference path.

8. The method according to claim 7, characterized in that, The intermediate reference path is stabilized to obtain the target reference path, including: Based on the intermediate reference path corresponding to the current time and the intermediate reference path corresponding to the previous time, determine the jump path point on the intermediate reference path corresponding to the previous time and the far path point on the intermediate reference path corresponding to the current time. The far path point is a path point on the intermediate reference path whose distance from the vehicle's location exceeds a preset distance threshold. Interpolation is performed between the transition path point and the remote path point to obtain the target reference path.

9. The method according to any one of claims 1-8, characterized in that, Before acquiring the trajectory information of the second vehicle ahead of the first vehicle at the intersection, the method further includes: Obtain the orientation angle of the third vehicle in front of the first vehicle relative to the centerline of the lane in which the third vehicle is located, and the lateral distance between the third vehicle and the centerline of the lane in which the first vehicle is located. Based on the navigation information of the first vehicle, the heading angle, and the lateral distance, a third vehicle with the same turning intention as the first vehicle is identified as the second vehicle.

10. The method according to claim 9, characterized in that, Based on the navigation information of the first vehicle, the heading angle, and the lateral distance, determining a third vehicle with the same turning intention as the first vehicle as the second vehicle includes: Based on the navigation information of the first vehicle, determine the turning intention of the first vehicle; Based on the stated steering intention, determine the target orientation angle range and the target lateral distance range; Based on the orientation angle and the target orientation angle range, as well as the lateral distance and the target lateral distance range, a second vehicle among the third vehicles with the same steering intention as the first vehicle is determined.

11. The method according to any one of claims 1-8, characterized in that, Before adjusting the initial reference path based on the trajectory information of the second vehicle to obtain the target reference path for trajectory planning, the method further includes: Obtain the initial trajectory information of the second vehicle; Based on the trajectory endpoint positions of the initial trajectory information, determine the completeness of the initial trajectory information; When the complete situation indicates that the initial trajectory information is incomplete, the initial trajectory information is completed to obtain the trajectory information.

12. A path planning device, characterized in that, include: The first acquisition module is configured to acquire trajectory information of a second vehicle ahead of the first vehicle in the intersection. The second acquisition module is configured to acquire an initial reference path corresponding to a target type lane in the road where the first vehicle is located, wherein the driving direction indicated by the target type lane is the same as the driving direction of the first vehicle at the intersection, and the initial reference path is used to connect the target type lane and the road opposite the target type lane at the intersection. The adjustment module is configured to adjust the initial reference path based on the trajectory information of the second vehicle to obtain a target reference path for trajectory planning.

13. A computer-readable storage medium having computer program instructions stored thereon, characterized in that, When the program instructions are executed by the processor, they implement the path planning method as described in any one of claims 1-11.

14. A vehicle, characterized in that, include: Storage device for storing computer programs; An execution device is used to execute the computer program to implement the path planning method according to any one of claims 1-11.

Images