A lane-level global path planning method

Abstract

This invention provides a lane-level global path planning method. First, high-precision map data is read, and based on the A*STAR algorithm, road-level path planning is performed from the starting point to the target point, generating a road-level path related to the driving distance, the number of intersections passed, and the number of turns. Finally, based on the road-level path planning, the road-level path is refined into lane-level paths, which can be divided into two steps: Step 1 selects lane-level connection nodes from the target point to the starting point based on the road-level planning results; Step 2 connects the planned paths from the starting point to the target point based on the lane-level connection nodes. The lane-level global path completed by this planning method has the advantage of early lane changing, giving autonomous vehicles sufficient lane-changing distance during actual driving.

Description

Technical Field

[0001] This invention belongs to the field of autonomous driving technology, and in particular relates to a lane-level global path planning method. Background Technology

[0002] Traditional road-level path planning is accurate enough for human drivers, but provides overly vague guidance for autonomous vehicles. To compensate for this, vehicles based on road-level navigation must possess powerful real-time environmental perception and decision-making systems to plan specific driving routes in real time, which significantly increases the burden on onboard computing units. With the development of autonomous driving technology, providing a lane-level global reference guidance trajectory that autonomous vehicles can follow, assuming no obstacles, is particularly important. This can greatly reduce the computational burden and lower the risk of system failure. Summary of the Invention

[0003] To address the technical problems existing in the background art, the present invention provides a lane-level global path planning method, which can quickly generate a lane-level global path planning method that considers multiple driving parameters and complies with traffic rules, providing a global reference guideline for autonomous vehicles.

[0004] The algorithm part of the lane-level global path planning method provided by this invention is based on the A* Star algorithm.

[0005] Specifically, this invention provides a lane-level global path planning method, comprising the following steps:

[0006] Step 1: Read high-precision map data and store the data in layers: road level, lane level, lane point level, and intersection connection relationship;

[0007] Step 2: Match the vehicle's starting point (StartPoint) and target point (EndPoint) with the map data to locate the vehicle at the lane level. If the match is successful, proceed to the next step; otherwise, the planning fails. During the matching process, the maximum matching distance between the vehicle's position and the lane in the high-precision map is defined as MaxDis. If the minimum distance from the vehicle's position to a nearby lane in the high-precision map is less than MaxDis, then all of these different lane points can be used as matching targets for road planning, and the minimum cost is selected as the final result.

[0008] Step 3: Create an Open and Close list, and put the vehicle's starting point into the Open list;

[0009] Step 4: Select the value with the minimum penalty cost from the Open list as the current node, remove it from the Open list and put it into the Close list; if the current node has the same road as the target point, proceed to step 6, otherwise proceed to step 5;

[0010] Step 5: Expand road-level nodes: Expand reachable nodes based on the current node. If a reachable node exists in Close, discard that node; if a reachable node does not exist in Close or Open, add it to Open. For nodes already existing in Open, select the node with the lowest penalty cost for node update; if Open is not empty, proceed to Step 4; otherwise, planning fails and the process ends.

[0011] Step 6: Extract the optimal road-level planning data RoadList from Close;

[0012] Step 7: Perform lane-level path connection. If the connection is successful, output the lane-level path planning result and the process ends; otherwise, modify the Close and Open lists; proceed to Step 5 and replan.

[0013] As a further explanation of the present invention, the penalty cost consists of road length, number of intersections passed, and turns.

[0014] As a further explanation of the present invention, the lane-level path connection process in step 7 includes the following steps:

[0015] Step 71: Create a WantRoadInf list and add the target point EndPoint;

[0016] Step 72: Select the newly added node in the WantRoadInf list as the CurrentNode;

[0017] Step 73: In RoadList, find the preceding connecting road of the road corresponding to CurrentNode, and find the shortest lane that can connect to the CurrentNode lane; if the CurrentNode lane cannot be directly connected to the preceding connecting road, find the WantLane lane in the road-level data corresponding to CurrentNode that can connect to the preceding connecting road, and connect the CurrentNode lane to the WantLane lane in the order of lane changes; if the connection is successful, put the connected lane information into the WantRoadInf list and proceed to Step 72; otherwise, return to road-level planning, output the truncated node, and proceed to Step 78; after the RoadList iteration is completed, proceed to Step 74.

[0018] Step 74: Create a list of RoadPoints and add the starting point StartPoint;

[0019] Step 75: Select the last node in the RoadPoint list as the CurrentNode;

[0020] Step 76: In WantRoadInf, find the next connecting lane of the lane corresponding to CurrentNode. If the lane corresponding to CurrentNode and the next connecting lane are connected, connect the lanes directly; otherwise, a lane-changing curve connection is required. If the lane-changing fails, search for a connecting road after CurrentNode based on the high-precision map in Step 1. If the connection is successful, add the connection point information to the RoadPoint list and proceed to Step 75; otherwise, return to road-level planning, output the truncated node, and proceed to Step 78; complete the search and reach the destination to proceed to Step 77.

[0021] Step 77: Output lane-level path planning results; process ends.

[0022] Step 78: Delete nodes at the same level as the truncated node in the Close list and add them to the Open list; or clear Close and Open, add all truncated nodes generated by the search of the start and end points to Close, and add the start point to Open; proceed to step 5 and replan.

[0023] As a further explanation of the present invention, when making the lane-changing curve connection, six parameters are defined: ① lane-changing distance after entering the intersection, ② lane-changing distance before entering the intersection, ③ lane-changing distance after entering the intersection, ④ the ratio of lane-changing distance after entering the intersection to the length of the lane, ⑤ the ratio of lane-changing distance before entering the intersection to the length of the lane, and ⑥ the ratio of lane-changing distance after entering the intersection to the length of the lane, so as to realize arbitrary adjustment of the lane-changing position.

[0024] Compared with the prior art, the present invention has the following beneficial technical effects:

[0025] This invention proposes a hierarchical lane-level global path planning method, which can quickly generate a lane-level global path planning algorithm that considers multiple driving parameters and complies with traffic rules, providing a global reference guideline for autonomous vehicles. The lane-level global path completed by the planning method provided by this invention has the advantage of early lane changing, giving autonomous vehicles sufficient lane-changing distance during actual driving. Attached Figure Description

[0026] Figure 1 The overall flowchart of the lane-level global path planning method provided by this invention;

[0027] Figure 2 The logic block diagram of the lane-level global path planning method provided by the present invention;

[0028] Figure 3 The lane-level path connection logic block diagram provided by this invention;

[0029] Figure 4 This is an example diagram of lane-level global path planning provided by the present invention. Detailed Implementation

[0030] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0031] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0032] In the description of this invention, it should also be noted that, unless otherwise explicitly specified and limited, the terms "set," "install," "connect," and "link" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0033] The technical solution of the present invention will be explained below with reference to specific embodiments.

[0034] Figure 2 The logic block diagram of the lane-level global path planning method provided by this invention is as follows: Figure 2 As shown, a lane-level global path planning method is provided, including the following steps:

[0035] Step 1: Read high-precision map data and store the data in layers such as road level, lane level, lane point level, and intersection connection relationship;

[0036] Step 2: Match the vehicle's starting point (StartPoint) and target point (EndPoint) with the map data to locate the vehicle at the lane level. If the match is successful, proceed to the next step; otherwise, the planning fails. During the matching process, the maximum matching distance between the vehicle's position and the lane in the high-precision map is defined as MaxDis. If the minimum distance from the vehicle's position to a nearby lane in the high-precision map is less than MaxDis, then all of these different lane points can be used as matching targets for road planning, and the minimum cost is selected as the final result.

[0037] Step 3: Create an Open and Close list, and put the vehicle's starting point into the Open list;

[0038] Step 4: Select the value with the minimum penalty cost from the Open list as the current node, remove it from the Open list and put it into the Close list; if the current node has the same road as the target point, proceed to step 6, otherwise proceed to step 5;

[0039] Step 5: Expand road-level nodes: Expand reachable nodes based on the current node. If a reachable node exists in Close, discard that node; if a reachable node does not exist in Close or Open, add it to Open. For nodes already existing in Open, select the node with the lowest penalty cost for node update; the penalty cost consists of road length, number of intersections passed, turning, etc.; if Open is not empty, proceed to step 4; otherwise, planning fails and the process ends.

[0040] Step 6: Extract the optimal road-level planning data RoadList from Close;

[0041] Step 7: Perform lane-level path connection. If the connection is successful, output the lane-level path planning result and the process ends; otherwise, modify the Close and Open lists; proceed to Step 5 and replan.

[0042] Figure 3 The lane-level path connection logic block diagram provided by this invention is as follows: Figure 3 As shown, the lane-level path connection process in step 7 above includes the following steps:

[0043] Step 71: Create a WantRoadInf list and add the target point EndPoint;

[0044] Step 72: Select the newly added node in the WantRoadInf list as the CurrentNode;

[0045] Step 73: In RoadList, find the preceding connecting road of the road corresponding to CurrentNode, and find the shortest lane that can connect to the CurrentNode lane; if the CurrentNode lane cannot be directly connected to the preceding connecting road, find the WantLane lane in the road-level data corresponding to CurrentNode that can connect to the preceding connecting road, and connect the CurrentNode lane to the WantLane lane in the order of lane changes; if the connection is successful, put the connected lane information into the WantRoadInf list and proceed to Step 72; otherwise, return to road-level planning, output the truncated node, and proceed to Step 78; after the RoadList iteration is completed, proceed to Step 74.

[0046] Step 74: Create a list of RoadPoints and add the starting point StartPoint;

[0047] Step 75: Select the last node in the RoadPoint list as the CurrentNode;

[0048] Step 76: In WantRoadInf, find the next connecting lane of the lane corresponding to CurrentNode. If the lane corresponding to CurrentNode and the next connecting lane are connected, connect the lanes directly; otherwise, a lane-changing curve connection is required. If the lane-changing fails, search for a connecting road after CurrentNode based on the high-precision map in Step 1. If the connection is successful, add the connection point information to the RoadPoint list and proceed to Step 75; otherwise, return to road-level planning, output the truncated node, and proceed to Step 78; complete the search and reach the destination to proceed to Step 77.

[0049] Step 77: Output lane-level path planning results; process ends.

[0050] Step 78: Delete nodes at the same level as the truncated node in the Close list and add them to the Open list; or clear Close and Open, add all truncated nodes generated by the search of the start and end points to Close, and add the start point to Open; proceed to step 5 and replan.

[0051] Preferably, when connecting the lane-changing curves, six parameters are defined: ① lane-changing distance after entering the intersection, ② lane-changing distance before entering the intersection, ③ lane-changing distance after entering the intersection, ④ the ratio of lane-changing distance after entering the intersection to the length of the lane, ⑤ the ratio of lane-changing distance before entering the intersection to the length of the lane, and ⑥ the ratio of lane-changing distance after entering the intersection to the length of the lane, so as to realize arbitrary adjustment of the lane-changing position.

[0052] Figure 4 The diagram illustrates a lane-level global path planning example provided by this invention. Start represents the starting point, and End represents the target point. As shown, a lane-level global path is generated from the starting point to the target point, and multiple lane changes are performed in advance while adhering to road driving rules.

[0053] The embodiments given above are preferred examples for implementing the present invention, and the present invention is not limited to the above embodiments. Any non-essential additions or substitutions made by those skilled in the art based on the technical features of the present invention are within the protection scope of the present invention.

Claims

1. A lane-level global path planning method, characterized in that, The method comprises the following steps: Step 1: reading high-precision map data, and storing the data in layers as road level-lane level-lane point level and intersection connection relationship; Step 2: matching the vehicle start point StartPoint and the target point EndPoint position with the map data, and positioning to the lane level of the map; if the matching is successful, entering the next step, otherwise, the planning fails; in the matching, the maximum matching distance of the vehicle position and the high-precision map lane is MaxDis, if the minimum distance from the vehicle position to the nearby lane in the high-precision map is less than MaxDis, then these different lane points can all be used as the matching target for road planning, and the minimum cost is selected as the final result; Step 3: creating an Open list and a Close list, and putting the vehicle start point into the Open list; Step 4: selecting the value with the minimum penalty cost in the Open list as the current node, removing the current node from the Open list and putting it into the Close list; if the current node is the same as the target point in the road, entering step 6, otherwise, entering step 5; Step 5: expanding the road level node: expanding the reachable node from the current node, if the reachable node exists in the Close list, abandoning the node; if the reachable node does not exist in the Close list, adding the node into the Open list, and if the node already exists in the Open list, updating the node by selecting the minimum penalty cost; if the Open list is not empty, entering step 4; otherwise, the planning fails, and the process ends; Step 6: extracting the best road level planning data RoadList from the Close list; Step 7: performing lane level path connection, if the connection is successful, outputting the lane level path planning result, and the process ends; otherwise, modifying the Close list and the Open list, entering step 5, and re-planning; The lane level path connection process of step 7 comprises the following steps: Step 71: creating a WantRoadInf list, and adding the target point EndPoint into the list; Step 72: selecting the newly added node in the WantRoadInf list as the CurrentNode; Step 73: finding the previous connection road of the CurrentNode in the RoadList, and finding the shortest lane that can connect the CurrentNode lane; if the CurrentNode lane cannot be directly connected to the previous connection road, finding the WantLane lane that can connect the previous connection road in the road level data corresponding to the CurrentNode, and connecting the CurrentNode lane to the WantLane lane according to the lane change order; if the connection is successful, putting the connection lane information into the WantRoadInf list, entering step 72; otherwise, returning to the road level planning, outputting the truncated node, entering step 78; after the iteration of the RoadList ends, entering step 74; Step 74: creating a RoadPoint list, and adding the start point StartPoint into the list; Step 75: selecting the last node in the RoadPoint list as the CurrentNode; Step 76: find the next connected lane of the lane corresponding to the CurrentNode in the WantRoadInf, if the lane corresponding to the CurrentNode and the next connected lane have a connection relationship, directly connect the lanes; otherwise, the lane changing curve connection is needed, if the lane changing fails, search the connectable lane of the next connected road of the CurrentNode according to step 1, if the connection is successful, put the connection point information into the RoadPoint list and enter step 75; otherwise, return to the road level planning, output the truncated node and enter step 78; complete the search and reach the end point to enter step 77; Step 77: output the lane level path planning result, and the flow ends; Step 78: delete the truncated node of the same layer node in the Close list and put it into the Open list, or clear the Close and Open, add all the truncated nodes generated by the search of the start point and the end point to the Close and Open, and add the start point to the Open; enter step 5 and plan again.

2. The lane-level global path planning method of claim 1, wherein, The penalty cost is composed of road length, number of passing through intersections and turning.

3. The lane-level global path planning method of claim 1, wherein, When performing the lane changing curve connection, six parameters are defined: ① lane changing distance after entering the intersection, ② lane changing distance before entering the intersection, ③ lane changing distance after lane changing, ④ ratio of lane changing distance after entering the intersection to the length of the lane, ⑤ ratio of lane changing distance before entering the intersection to the length of the lane, and ⑥ ratio of lane changing distance after lane changing to the length of the lane, so as to realize arbitrary adjustment of the lane changing position.

Images