Search

CN-122015883-A - Parking path planning method and device, storage medium, vehicle-mounted equipment and vehicle

CN122015883ACN 122015883 ACN122015883 ACN 122015883ACN-122015883-A

Abstract

The application discloses a parking path planning method, a device, a storage medium, vehicle-mounted equipment and a vehicle. The parking path planning method comprises the steps of identifying a turning point candidate area, selecting a target turning point from the turning point candidate area, searching a first effective path from a vehicle position point to the target turning point based on a narrow channel parameter, searching a second effective path from the target turning point to a vehicle self-point based on a broken road parameter, and carrying out path fusion on the first effective path and the second effective path to obtain a parking path. By the method, high-precision and high-efficiency path planning under a narrow-channel and broken-end combined scene is realized, the success rate, searching efficiency and robustness of path planning can be remarkably improved, and meanwhile, the method has stronger scene adaptability and expandability, and is beneficial to promoting the floor application of intelligent driving vehicles in the narrow-channel and broken-end combined scene.

Inventors

  • FU CHENGYU

Assignees

  • 深圳元戎启行科技有限公司

Dates

Publication Date
20260512
Application Date
20251223

Claims (10)

  1. 1. A method of parking path planning, comprising: Identifying a turning point candidate region; selecting a target turning point from the turning point candidate area; Searching a first effective path from a vehicle position point to the target turning point based on the narrow channel parameters; searching a second effective path from the target turning point to the vehicle point based on the broken road parameter; and carrying out path fusion on the first effective path and the second effective path to obtain a parking path.
  2. 2. The parking path planning method according to claim 1, characterized in that the identifying a turning point candidate region includes: And identifying the area of which the area in the preset driving range is not smaller than the sum value of the minimum turning space area and the safety redundant area of the vehicle and the area of which the distance between the area and each obstacle is not smaller than the maximum profile radius of the vehicle as the turning point candidate area.
  3. 3. The parking path planning method according to claim 1, wherein the selecting a target turning point from the turning point candidate region includes: Screening effective turning points in the turning point candidate area based on a preset kinematic constraint index and a preset accessibility index; and scoring each effective turning point based on the turning space adequacy index and the parking space distance adjacency index, and selecting the effective turning point with the highest score as the target turning point.
  4. 4. The parking path planning method according to claim 1, wherein the searching for the first effective path from the vehicle location to the target u-turn point based on the narrow-channel parameter includes: Searching a plurality of nodes with the lowest priority cost from the vehicle locus to the target turning point for obstacle avoidance based on the narrow channel parameters, and obtaining the first effective path; The obstacle avoidance priority cost comprises a first actual cost and a first heuristic cost, the first actual cost comprises a steering angle cost and a narrow channel boundary distance punishment cost from the vehicle locus to the corresponding node, and the first heuristic cost comprises a motion curve length punishment cost and a gear shifting frequency cost from the corresponding node to the target turning point.
  5. 5. The parking path planning method according to claim 4, wherein the searching for a second effective path from the target turning point to the own vehicle point based on the break path parameter includes: searching a plurality of nodes with lowest priority cost from the target turning point to the self-driving point based on the broken road parameters to obtain the second effective path; The efficiency priority cost comprises a second actual cost and a second heuristic cost, the second actual cost comprises a path length cost and a steering smoothness punishment cost from the target turning point to the corresponding node, and the second heuristic cost comprises a motion curve length punishment cost and a gear shifting frequency cost from the corresponding node to the self-driving point.
  6. 6. The method for planning a parking path according to claim 1, wherein the performing path fusion on the first effective path and the second effective path to obtain a parking path includes: Smoothing the first effective path, the second effective path and the path connecting section by adopting a B spline curve to obtain the parking path; And the smooth processing restricts the continuity index of the path connection section adaptive turning and steering and restricts the maximum curvature of the parking path not to be larger than the reciprocal of the minimum turning radius of the vehicle.
  7. 7. A parking path planning apparatus, comprising: the turning point preprocessing module is used for identifying a turning point candidate area and selecting a target turning point from the turning point candidate area; The hierarchical path searching module is used for searching a first effective path from a vehicle position point to the target turning point based on the narrow channel parameters and searching a second effective path from the target turning point to a vehicle self point based on the broken road parameters; And the path fusion module is used for carrying out path fusion on the first effective path and the second effective path to obtain a parking path.
  8. 8. A storage medium having stored thereon program data, wherein the program data, when executed by a processor, implements the steps of the parking path planning method according to any one of claims 1 to 6.
  9. 9. A vehicle-mounted device comprising a processor and a memory connected to each other, the memory storing a computer program, the processor, when executing the computer program, implementing the steps of the parking path planning method according to any one of claims 1 to 6.
  10. 10. A vehicle characterized in that the vehicle includes at least one of the parking path planning apparatus according to claim 7, the storage medium according to claim 8, and the in-vehicle apparatus according to claim 9.

Description

Parking path planning method and device, storage medium, vehicle-mounted equipment and vehicle Technical Field The application relates to the technical field of intelligent driving, in particular to a parking path planning method, a device, a storage medium, vehicle-mounted equipment and a vehicle. Background In an intelligent driving parking scene, the planning of a parking path is one of the key links for realizing autonomous parking of a vehicle, and a parking track which meets the kinematic constraint of the vehicle, the obstacle avoidance safety requirement and has smooth and continuous path needs to be generated, so that the efficiency and the safety of the parking process are directly influenced. At present, the A-search algorithm is widely applied to parking path planning due to the characteristics of high search efficiency, strong path controllability and the like. In the path planning process of different parking scenes, the algorithm needs to adapt to scene requirements by setting specific search parameters so as to balance the feasibility and the search efficiency of the path, for example, in a narrow-channel scene, smaller grid size and step length are selected to fully utilize space to obtain a feasible path so as to ensure the accuracy of path planning, and in a broken-end scene, larger grid size and step length are selected to quickly generate a turning track so as to improve the search speed. Meanwhile, in the path searching process, the adopted path Cost function Cost (n) is generally defined as the sum of the actual Cost g (n) from the starting point to the current node and the heuristic Cost h (n) from the current node to the target point, wherein the heuristic Cost needs to satisfy the constraint of less than or equal to the real optimal Cost (Cost-to-Go) so as to ensure the optimality of the searching result. However, in a complex parking scenario where a narrow channel and a broken road are combined (as illustrated in fig. 1, the channel width is only 30-50cm greater than the length of a vehicle, the end of the channel is connected with a target parking space, the end of the channel is a wall or fixed obstacle to form a broken-end scenario, a trapezoid area protruding in the vehicle graph is used for providing a u-turn space, and particularly referring to the following description of fig. 3, a forward arrow indicates the initial state of the vehicle before entering the combined scenario), the prior art has significant limitations that firstly, the dual requirements of the combined scenario are difficult to be met by a single parameter, if fine grid modeling is adopted, the searching speed is possibly excessively slow, and if coarse grid modeling is adopted, an accurate feasible path is difficult to be generated near the entrance of the parking space, secondly, the heuristic function design is difficult to effectively guide the vehicle to find a reasonable path, the vehicle is easily sunk into the surrounding area to be the u-turn around, and then enters the cost, thirdly, the cost is often lower than the cost when the vehicle is driven back to the surrounding area near the obstacle, and the optimal path is far lower than the actual node, and the heuristic cost is greatly expanded near the optimal node, and the optimal path is greatly reduced, and the heuristic path is not expanded in the whole direction is greatly improved. These problems seriously affect the robustness and instantaneity of the parking path planning of the intelligent driving vehicle in a complex limited scene, and the path feasibility is poor. Disclosure of Invention The application mainly provides a parking path planning method, a device, a storage medium, vehicle-mounted equipment and a vehicle, and aims to solve the problem that the path feasibility of the traditional parking path planning mode is poor in a narrow-channel and broken-end combined scene. In order to solve the technical problems, the technical scheme adopted by the application is to provide a parking path planning method. The parking path planning method comprises the steps of identifying a turning point candidate area, selecting a target turning point from the turning point candidate area, searching a first effective path from a vehicle position point to the target turning point based on a narrow channel parameter, searching a second effective path from the target turning point to a vehicle self-point based on a broken-head path parameter, and carrying out path fusion on the first effective path and the second effective path to obtain a parking path. In some embodiments, the identifying the turning point candidate region includes identifying a region in which an area within a preset driving range is not smaller than a sum of a minimum turning space area of the vehicle and a safety redundant area, and a distance between the region and each obstacle is not smaller than a maximum contour radius of the vehicle as the turning point candidate region. In s