Lane-level global path planning method and device based on high-precision port map

Abstract

The application discloses a lane-level global path planning method and device based on a high-precision map of a port, and mainly has the following technical scheme: a static high-precision map is loaded. Based on the port environment, the static features of the road environment are acquired through a map module, and the static high-precision map is output; a landing and unloading bridge template is added. On the basis of the static high-precision map, according to the position of the berthing of a ship, the landing and unloading bridge templates are added at the bow and the stern of the ship respectively; a work-related high-precision map is generated. The static high-precision map and the landing and unloading bridge template vectors are superimposed and the lane topological relationship is reconstructed, and the work-related high-precision map of the unmanned vehicle is generated, so as to provide the environmental positioning information for subsequent path planning. A path planning graph is constructed. Based on the work-related high-precision map of the unmanned vehicle, the lane information between the starting point of the task and the ending point of the task is searched. Global path dynamic search is performed.

Description

Technical Field

[0001] This invention relates to the field of autonomous driving technology, and in particular to a lane-level global path planning method and apparatus based on a high-precision port map. Background Technology

[0002] In recent years, the construction and development of fully unmanned and fully automated ports has progressed rapidly. In closed or semi-open port environments, driverless container trucks or driverless container flatbed trucks (collectively referred to as driverless vehicles) are gradually replacing manually driven vehicles in ports. Global path planning is a crucial part of the entire driverless system; safe and efficient path planning can better ensure the smooth completion of port operations. To ensure the accuracy and reliability of global path planning, the vehicle needs real-time and accurate positioning of environmental information and high-precision maps of the operating environment. However, due to the special nature of port driverless scenarios, there are still certain limitations in global path planning for large-scale, two-way lane driving environments.

[0003] The unmanned driving scenario in ports differs from that in urban or highway scenarios, mainly in the following aspects: (1) The road under the quay crane (quay crane road) has a two-way structure, meaning that in the port road environment, the lane direction can be switched at any time according to the direction of the ship's berthing; (2) The path for going up and down the quay crane is arbitrary. That is, there is usually no lane line between the quay crane road and the latitude road. The internal port road parallel to the quay crane road is generally called the latitude road. The process from the latitude road to the quay crane road is defined as going up the quay crane, and the process from the quay crane road to the latitude road is defined as going down the quay crane. In principle, one can go up and down the quay crane at any point on the quay crane road. But in reality, one usually goes up the quay crane near the stern and goes down the quay crane near the bow; (3) The lock area of ​​the container loading and unloading lock is uncertain. It is necessary to decide whether to go through the lock area of ​​the container loading and unloading lock according to the port operation requirements. Therefore, although a lot of research has been carried out on the global path planning and tracking of unmanned vehicles and other robots in urban or highway environments, it is not easy to directly apply these methods to unmanned vehicles in port operation environments.

[0004] Currently, common methods in global path planning technology for autonomous vehicles include genetic algorithms, visual graph methods, artificial potential field methods, and others. These methods can all achieve path planning. However, global planning for autonomous driving is a complex problem. Vehicles must not only comply with various traffic rules, but also cope with dynamic traffic environments, avoid collisions, and pursue driving efficiency and quality. Path planning methods have been widely applied to path planning for unstructured road robots or autonomous vehicles, and their flexibility and real-time characteristics meet the obstacle avoidance requirements of autonomous driving. Meanwhile, in port operation environments, rich road environment features are required; therefore, it is necessary to provide a lane-level global path planning method and device based on high-precision port maps to achieve safe and efficient path planning. Summary of the Invention

[0005] In view of this, the present invention provides a lane-level global path planning method and apparatus based on a high-precision port map. The main purpose is to provide a safe and efficient lane-level global path planning method in a port environment, focusing on solving problems such as two-way lanes of quay cranes, difficulties in getting on and off quay cranes, and dynamically deciding whether to pass through container loading and unloading lock areas.

[0006] To achieve the above objectives, the present invention mainly provides the following technical solutions:

[0007] On one hand, this invention provides a lane-level global path planning method based on a high-precision port map, the method comprising:

[0008] S1, Load a static high-precision map. Based on the port environment, acquire static features of the road environment (including lane lines, road signs, etc.) through the map module, and output a static high-precision map of the port environment;

[0009] S2, Add templates for dock access bridges. Based on the static high-precision port environment map, add dock access bridge templates at the bow and stern of the ship according to the berthing position;

[0010] S3. Generate a high-precision map related to the operation. The static high-precision map is overlaid with the template vectors of the upper and lower shore bridges, and the lane topology is reconstructed to generate a high-precision map related to the operation of the unmanned vehicle, providing environmental positioning information for subsequent path planning.

[0011] S4, Construct a path planning map. This is done using high-precision maps relevant to the autonomous vehicle's operations. Lane nodes in Topological relationship between lanes The relationships between them are as follows:

[0012]

[0013] S5, Locate the lanes of the start and end points. Based on high-precision maps related to the autonomous vehicle operation, search for lane information between the task start and end points.

[0014] S6, global path dynamic search. Based on classic autonomous vehicle technology. The algorithm has been improved for port scenarios, enabling it to handle two-way lane planning problems and determine during planning whether to prohibit lane changes for a certain lane_node or whether to allow passage through container loading / unloading lock areas. Specifically, after the improvement... The algorithm is as follows:

[0015]

[0016] in, It is the starting point To the current node The minimum practical cost.

[0017]

[0018] in, From the starting point To the current node All possible paths; yes One of the paths; yes Lane junctions passed . The current node The estimated cost.

[0019] To ensure the improvement The algorithm can find the optimal global path. Cannot exceed the node To the finish line The actual cost. Therefore, choose To the finish line The Euclidean distance, i.e.:

[0020]

[0021] Optionally, the path planning map is constructed using a high-precision map related to the operation of the unmanned vehicle. Lane nodes in Topological relationship between lanes ,include:

[0022] S41, Lane Node Definition and selection. Specifically,

[0023] Each lane node is associated with the topological edges between the preceding lanes, defined as `predecessor_edges`, the topological edges between the succeeding lanes, defined as `successor_edges`, the topological edges between the left adjacent lanes, defined as `left_edges`, and the topological edges between the right adjacent lanes, defined as `right_edges`.

[0024] Among them, `predecessor_edge` points to a lane node in the direction of travel of the path, and there may be multiple such nodes. `successor_edge` points to the next lane node in the direction of travel of the path, and there may be multiple such nodes.

[0025] left_edge: Points to the left adjacent lane node.

[0026] right_edge: Points to the right adjacent lane node.

[0027] S42, probability determination of lane nodes. Specifically,

[0028] There are two possibilities for a lane node not having left_edges between its left and adjacent lanes: (1) the leftmost lane; (2) lane changing to the left is prohibited.

[0029] Similarly, a lane node without right_edges between its right adjacent lanes may have two possibilities: (1) the rightmost lane; (2) the right loop is prohibited.

[0030] The lane node pointed to by the successor_edge in the topological relationship between subsequent lanes is called the successor, the lane node pointed to by the left_edge in the topological relationship between the left adjacent lanes is called the left_neighbor, and the lane node pointed to by the right_edge in the topological relationship between the right adjacent lanes is called the right_neighbor. The three are collectively referred to as neighbors.

[0031] Two-way lanes have both positive and negative neighbors, while one-way lanes only have positive neighbors.

[0032] S43, Definition and cost calculation of topological edges. Specifically,

[0033] Each topological edge contains a starting lane node (lane_node), an ending lane node (lane_node), and a toll cost (Call).

[0034] Call = Cbase + Cadd

[0035] Cbase is the base cost, which is proportional to the lane length; Cadd is the additional cost, which can be the cost of changing lanes to the left, changing lanes to the right, turning left, turning right, making a U-turn, etc., and can be dynamically adjusted.

[0036] Optionally, the step of searching for lane information between the task start point and the task end point based on the high-precision map related to the unmanned vehicle operation includes:

[0037] S51, extract waypoints from the high-precision map related to the task. Specifically,

[0038] Generally, waypoints are extracted at equal intervals, for example, one waypoint every 0.5 meters. Important operational task points, such as yard bay locations, quay crane berth locations, lockout locations, charging pile locations, and parking locations, can also be extracted as a single waypoint.

[0039] S52, construct the KD-Tree of waypoints, and output the data structure of waypoints in k-dimensional Euclidean space.

[0040] S53, in the KD-Tree structure, find the nearest neighbor waypoints of the task start and end points.

[0041] S54, find the lane node corresponding to the nearest roadpoint (waypoint) between the task start and end points.

[0042] Optionally, for port scenarios, this is based on improvements made to unmanned vehicles. The algorithm performs a global path dynamic search, including:

[0043] S61, Definition and Judgment of Dynamic Cost. Specifically,

[0044] Lane changing is prohibited during planning: prohibiting left lane changing sets the additional cost Cadd of the left adjacent edge to infinity ∞; prohibiting right lane changing sets the additional cost Cadd of the right adjacent edge to infinity ∞.

[0045] During planning, decide whether to pass through the lock zone: If it is decided not to pass through the container loading and unloading lock zone based on the task, set the additional cost Cadd of the successor edge of the lane_node to which the lock zone belongs to infinity ∞.

[0046] S62, the search order and planning method for two-way lanes are defined. Specifically,

[0047] Positive planning: The algorithm searches for positive neighbors in the current lane_node.

[0048] positive_neighbors={successors,left_neighbors,right_neighbors}

[0049] Reverse programming: The algorithm searches from the negative_neighbors of the current lane_node.

[0050] negative_neighbors={predecessors,left_neighbors,right_neighbors}

[0051] On the other hand, the present invention also provides a lane-level global path planning device based on a high-precision port map, the device comprising:

[0052] The static high-precision map loading unit is used to acquire static features of the port road environment (including lane lines, road signs, etc.) through the map module and output a static high-precision map of the port environment.

[0053] The module adding unit for berth access bridges is used to add templates for berth access bridges at the bow and stern of a ship, respectively, based on a static high-precision port environment map and according to the ship's berthing position.

[0054] The operation-related high-precision map generation unit is used to overlay static high-precision maps with the template vectors of the upper and lower shore bridges and reconstruct the lane topology to generate an operation-related high-precision map for unmanned vehicles, providing environmental positioning information for subsequent path planning.

[0055] The path planning graph construction unit is used to construct lane nodes and topological edges in the routing graph using high-precision maps related to autonomous vehicle operations.

[0056] The lane lookup unit is used to search for lane information between the start and end points of a task based on a high-precision map related to the operation of the unmanned vehicle.

[0057] A global path dynamic search unit for improving autonomous vehicle-based systems. The algorithm handles two-way lane planning problems and determines whether to prohibit a lane_node from changing lanes or whether to allow passage through container loading / unloading lock zones during planning.

[0058] Optionally, the path planning graph construction unit includes:

[0059] The lane node definition module is used to define the topological relationship edges between each lane node and the preceding lane, the following lane, the left adjacent lane, and the right adjacent lane.

[0060] The lane node determination module is used to determine whether each lane node has a corresponding topological relationship edge between the preceding lane, the following lane, the left adjacent lane, and the right adjacent lane, and the probability of its existence.

[0061] The topology edge definition module is used to define each topology edge, including the starting lane node, the ending lane node, and the passage cost.

[0062] Optionally, the lane location unit for finding the origin and destination includes:

[0063] The high-precision map waypoint extraction module is used to determine waypoint extraction rules, generally extracting waypoints at equal intervals, such as one waypoint every 0.5 meters. Important operational task points, such as yard berth locations, quay crane berth locations, lockout locations, charging pile locations, and parking locations, can also be extracted as waypoints.

[0064] The KD-Tree building block for waypoints is used to output the data structure of waypoints in k-dimensional Euclidean space;

[0065] The nearest neighbor path search module is used to find the nearest neighbor path between the task start point and the task end point in the KD-Tree structure.

[0066] The lane node lookup module is used to find the lane nodes corresponding to the nearest neighbor road points of the task start and task end points.

[0067] Optionally, the global path dynamic search unit includes:

[0068] The dynamic cost definition and judgment module is used to define dynamic cost selection values. For example, when planning to prohibit lane changes: if left lane changes are prohibited, the additional cost Cadd of the left adjacent edge is set to infinity (∞); if right lane changes are prohibited, the additional cost Cadd of the right adjacent edge is set to infinity (∞). When planning to decide whether to pass through a lock zone: if the task determines not to pass through the container loading / unloading lock zone, the additional cost Cadd of the successor edge of the lane_node to which the lock zone belongs is set to infinity (∞).

[0069] The two-way lane planning method determination module is used to determine the forward and reverse planning processes. Algorithm search order.

[0070] In another aspect, the present invention also provides a storage medium comprising a stored program, wherein, when the program is executed, the device on which the storage medium is located executes the lane-level global path planning method based on a high-precision port map as described above.

[0071] In another aspect, the present invention also provides an electronic device, the device comprising at least one processor, and at least one memory and bus connected to the processor;

[0072] The processor and the memory communicate with each other via the bus.

[0073] The processor is used to call program instructions in the memory to execute the lane-level global path planning method based on the high-precision port map as described above.

[0074] By employing the above-described technical solution, the technical solution provided by the present invention has at least the following advantages:

[0075] This invention provides a lane-level global path planning method and apparatus based on a high-precision port map. The invention generates an operation-related high-precision map by loading a static high-precision map of the port environment and adding on / off quay crane templates. Furthermore, it constructs a path planning map, locates the lanes containing the start and end points, and utilizes an improved approach based on unmanned vehicles. The algorithm enables dynamic global path search. This invention can provide rich road environment features in port operation environments, especially solving problems such as two-way lanes for quay cranes, difficulties in getting on and off quay cranes, and dynamically deciding whether to pass through container loading and unloading lock areas, ultimately achieving safe and efficient path planning.

[0076] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and in order to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0077] Figure 1 A flowchart of a lane-level global path planning method based on a high-precision port map is provided in this embodiment of the invention.

[0078] Figure 2 This is a schematic diagram illustrating the addition of a template for a quay bridge based on a high-precision port map, as provided in an embodiment of the present invention.

[0079] Figure 3 This invention provides a schematic diagram of the lane before adding templates to the formwork of an inlet / outlet bridge.

[0080] Figure 4 This is a schematic diagram of the lane after adding templates to the formwork of an inlet / outlet bridge, provided in an embodiment of the present invention.

[0081] Figure 5 A high-precision map related to operations—a schematic diagram of a two-way shore bridge road—is provided as an embodiment of the present invention.

[0082] Figure 6 A logical diagram of lane nodes and the topological relationship between lanes is provided in this embodiment of the invention.

[0083] Figure 7 A block diagram of another lane-level global path planning device based on a high-precision port map provided in an embodiment of the present invention;

[0084] Figure 8 An electronic device for lane-level global path planning of a high-precision port map provided in this embodiment of the invention. Detailed Implementation

[0085] Exemplary embodiments of the invention will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the invention are shown in the drawings, it should be understood that the invention may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this invention will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art.

[0086] This invention provides a lane-level global path planning method based on a high-precision port map, such as... Figure 1 As shown, this method generates a high-precision map related to operations by loading a static high-precision map of the port environment and adding templates for berth access bridges. Further, it constructs a path planning map, identifies the lanes containing the start and end points, and utilizes improved autonomous vehicle technology. The algorithm implements dynamic global path search, and the specific steps provided in this embodiment of the invention are as follows:

[0087] 101. Load a static high-precision map. Based on the port environment, acquire static features of the road environment (including lane lines, road signs, etc.) through the map module, and output a static high-precision map of the port environment;

[0088] 102. Add templates for dock access bridges. Based on the static high-precision port environment map, add templates for dock access bridges at the bow and stern according to the berthing positions of the ships;

[0089] 103. Generate high-precision maps related to the operation. Overlay the static high-precision map with the template vectors of the upper and lower shore bridges and reconstruct the lane topology to generate high-precision maps related to the operation of unmanned vehicles, providing environmental positioning information for subsequent path planning.

[0090] 104. Construct a path planning map. Utilize high-precision maps relevant to the unmanned vehicle's operations to construct... Lane nodes in Topological relationship between lanes .

[0091] In this embodiment of the invention, the mutual relationship is defined as follows:

[0092]

[0093] 105. Locate the lanes of the start and end points. Based on the high-precision map related to the unmanned vehicle operation, search for lane information between the task start and end points.

[0094] 106. Dynamic Global Path Search. Based on classic autonomous vehicle technology. The algorithm has been improved for port scenarios, enabling it to handle two-way lane planning problems and determine during planning whether to prohibit lane changes for a certain lane_node or whether to allow passage through container loading / unloading lock areas. Specifically, after the improvement... The algorithm is as follows:

[0095]

[0096] in, It is the starting point To the current node The minimum practical cost;

[0097]

[0098] in, From the starting point To the current node All possible paths; yes One of the paths; yes Lane junctions passed ; The current node The estimated cost; It is the cost of passage; Basic cost; As an additional cost.

[0099] In this embodiment of the invention, to ensure the improved The algorithm can find the optimal global path. Cannot exceed the node To the finish line The actual cost. Therefore, choose To the finish line The Euclidean distance, i.e.:

[0100]

[0101] To provide a more detailed explanation of the above embodiments, this invention also provides a method for adding templates for quay cranes based on high-precision port maps, such as... Figure 2 As shown, the method provides a detailed description of the above embodiments and also adds an illustration of the effect of adding templates for the berth bridge, that is, adding templates for the berth bridge at the bow and stern of the ship according to the berthing position of the ship.

[0102] Furthermore, to supplement the comparative explanation of the effect after adding the port access bridge template based on the high-precision port map, and to more specifically explain the process of vector overlaying the static high-precision map and the port access bridge template to reconstruct the topology and generate a work-related high-precision map, this embodiment of the invention also provides a comparative diagram of the lane layout before and after adding the port access bridge template. For example... Figure 3 and Figure 4 As shown in the figure. In addition, this embodiment of the invention also provides the following specific methods for representing lane topology relationships, as shown in Tables 1 and 2, which respectively represent the lane topology relationships before and after adding the template.

[0103] Table 1. Lane topology before adding template

[0104]

[0105] Table 2 Lane topology after adding templates

[0106]

[0107]

[0108] In this embodiment of the invention, the application scenario is as follows: A static high-precision map of the port environment is pre-obtained; an unmanned vehicle travels within the port; the quay crane environment is a two-way lane system; that is, it is necessary to construct lane nodes (lane_nodes) and topological relationships (topological_edges) between lanes using a high-precision map related to the operation. For example, obtaining... Figure 5 The image shows a high-precision map related to the operation - a schematic diagram of the two-way shore bridge road.

[0109] In this embodiment of the invention, a path planning map is constructed using high-precision maps related to the operation of unmanned vehicles. Lane nodes in Topological relationship between lanes ,include:

[0110] First, lane nodes ( The definition and selection of ). Specifically,

[0111] Each lane node is associated with the topological edges between the preceding lanes, defined as `predecessor_edges`, the topological edges between the succeeding lanes, defined as `successor_edges`, the topological edges between the left adjacent lanes, defined as `left_edges`, and the topological edges between the right adjacent lanes, defined as `right_edges`.

[0112] Among them, `predecessor_edge` points to a lane node in the direction of travel of the path, and there may be multiple such nodes. `successor_edge` points to the next lane node in the direction of travel of the path, and there may be multiple such nodes.

[0113] left_edge: Points to the left adjacent lane node.

[0114] right_edge: Points to the right adjacent lane node.

[0115] Secondly, the probability determination of lane nodes (lane_nodes), such as Figure 6 shown. Specifically,

[0116] There are two possibilities for a lane node not having left_edges between its left and adjacent lanes: (1) the leftmost lane; (2) lane changing to the left is prohibited.

[0117] Similarly, a lane node without right_edges between its right adjacent lanes may have two possibilities: (1) the rightmost lane; (2) the right loop is prohibited.

[0118] The lane node pointed to by the successor_edge in the topological relationship between subsequent lanes is called the successor, the lane node pointed to by the left_edge in the topological relationship between the left adjacent lanes is called the left_neighbor, and the lane node pointed to by the right_edge in the topological relationship between the right adjacent lanes is called the right_neighbor. The three are collectively referred to as neighbors.

[0119] Two-way lanes have both positive and negative neighbors, while one-way lanes only have positive neighbors.

[0120] Finally, the definition and cost calculation of topological edges are discussed. Specifically,

[0121] Each topological edge contains a starting lane node (lane_node), an ending lane node (lane_node), and a toll cost (Call).

[0122] Call = Cbase + Cadd

[0123] Cbase is the base cost, which is proportional to the lane length; Cadd is the additional cost, which can be the cost of changing lanes to the left, changing lanes to the right, turning left, turning right, making a U-turn, etc., and can be dynamically adjusted.

[0124] Furthermore, as a response to the above Figure 1 The present invention provides a lane-level global path planning device based on a high-precision port map to implement the method shown. This device embodiment corresponds to the aforementioned method embodiment. For ease of reading, this device embodiment will not repeat the details of the aforementioned method embodiment, but it should be understood that the device in this embodiment can implement all the contents of the aforementioned method embodiment. This device is applied to implement lane-level global path planning within a port, specifically as follows... Figure 7 As shown, the device 70 includes:

[0125] The static high-precision map loading unit 701 is used to acquire static features of the port road environment (including lane lines, road signs, etc.) through the map module and output a static high-precision map of the port environment.

[0126] The module addition unit 702 for the berth bridge is used to add templates for berth bridges at the bow and stern of a ship based on the static high-precision port environment map and the ship's berthing position.

[0127] The high-precision map generation unit 703 is used to overlay the static high-precision map with the template vectors of the upper and lower shore bridges and reconstruct the lane topology to generate a high-precision map related to the operation of the unmanned vehicle, providing environmental positioning information for subsequent path planning.

[0128] The path planning graph construction unit 704 is used to construct lane nodes and topological edges in the routing graph using high-precision maps related to unmanned vehicle operations.

[0129] The lane search unit 705, which is used to search for lane information between the start and end points of a task based on a high-precision map related to the operation of the unmanned vehicle.

[0130] Global path dynamic search unit 706, used for improvements based on autonomous vehicles. The algorithm handles two-way lane planning problems and determines whether to prohibit a lane_node from changing lanes or whether to allow passage through container loading / unloading lock zones during planning.

[0131] In summary, this invention provides a lane-level global path planning method and apparatus based on a high-precision port map. This invention generates a high-precision map related to operations by loading a static high-precision map of the port environment and adding on / off quay crane templates. Furthermore, by constructing a path planning map and finding the lanes containing the start and end points, it utilizes an improved approach based on unmanned vehicles. The algorithm enables dynamic global path search. This invention provides rich road environment features in port operations, particularly addressing issues such as two-way lanes for quay cranes, difficulties in accessing and exiting quay cranes, and dynamically deciding whether to pass through container loading / unloading lock areas, ultimately achieving safe and efficient path planning. Furthermore, the method provided in this invention is effectively applicable to complex port environments, particularly in two-lane and multi-lane areas such as quay cranes and storage yards, solving problems inherent in existing technologies and expanding the scope of port operations.

[0132] The lane-level global path planning device based on a high-precision port map includes a processor and a memory. The aforementioned fusion positioning unit, point cloud conversion unit, feature extraction unit, and map building unit are all stored as program units in the memory. The processor executes the aforementioned program units stored in the memory to achieve the corresponding functions.

[0133] The processor contains a kernel, which retrieves the corresponding program units from memory. One or more kernels can be configured, and adjusting kernel parameters provides an effective method for constructing lane-level global path planning within the port, ensuring the accuracy and quality of global path planning and guaranteeing the safe operation of autonomous vehicles.

[0134] This invention provides a storage medium storing a program that, when executed by a processor, implements the lane-level global path planning method based on a high-precision port map.

[0135] This invention provides a processor for running a program, wherein the program executes the lane-level global path planning method based on a high-precision port map.

[0136] This invention provides an electronic device 80, such as... Figure 8As shown, the device includes at least one processor 801, and at least one memory 802 and bus 803 connected to the processor 801; wherein the processor 801 and the memory 802 communicate with each other through the bus 803; the processor 801 is used to call program instructions in the memory 802 to execute the above-mentioned lane-level global path planning method based on port high-precision map.

[0137] The devices mentioned in this article can be servers, PCs, tablets, mobile phones, etc.

[0138] This application also provides a computer program product, which, when executed on a data processing device, is suitable for executing an initialization program with the following method steps: generating an operation-related high-precision map by loading a static high-precision map of the port environment and adding on / off quay crane templates. Further, by constructing a path planning map, finding the lanes where the start and end points are located, and based on improvements made by unmanned vehicles... The algorithm enables dynamic global path search.

[0139] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0140] In a typical configuration, the device includes one or more processors (CPUs), memory, and a bus. The device may also include input / output interfaces, network interfaces, etc.

[0141] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, like read-only memory (ROM) or flash RAM, and memory includes at least one memory chip. Memory is an example of computer-readable media.

[0142] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0143] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0144] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0145] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A lane-level global path planning method based on a high-precision port map, characterized in that, The method includes: Load a static high-precision map; based on the port environment, obtain static features of the road environment through the map module, and output a static high-precision map of the port environment; wherein, the static features of the road environment include lane lines and road signs; Add templates for dock access bridges; based on the static high-precision port environment map, add templates for dock access bridges at the bow and stern of the ship according to the berthing position; Generate a high-precision map related to the operation; overlay the static high-precision map with the template vectors of the upper and lower shore bridges and reconstruct the lane topology to generate a high-precision map related to the operation of the unmanned vehicle, providing environmental positioning information for subsequent path planning; Construct a path planning map; utilize high-precision maps related to the operation of unmanned vehicles to construct... Lane nodes in Topological relationship between lanes The relationships between them are as follows: Locate the lanes where the start and end points are located; based on high-precision maps related to the operation of unmanned vehicles, search for lane information between the task start and end points; Global path dynamic search; based on classic autonomous vehicle technology The algorithm has been improved for port scenarios, enabling it to handle two-way lane planning problems and determine during planning whether to prohibit lane changes for a certain lane_node or whether to allow passage through container loading and unloading lock areas; the improved algorithm... The algorithm is as follows: in, It is the starting point To the current node The minimum practical cost; in, From the starting point To the current node All possible paths; yes One of the paths; yes Lane junctions passed ; The current node The estimated cost; It is the cost of passage; Basic cost; As an additional cost; To ensure the improvement The algorithm can find the optimal global path. Cannot exceed the node To the finish line The actual cost; therefore, choose To the finish line The Euclidean distance, i.e.: 。 2. The method according to claim 1, characterized in that, The construction path planning map is built using high-precision maps related to the operation of unmanned vehicles. Lane nodes in Topological relationship between lanes ,include: Lane node Definition and selection Each lane node is associated with the topological edges between the preceding lanes, defined as `predecessor_edges`; the topological edges between the succeeding lanes, defined as `successor_edges`; the topological edges between the left adjacent lanes, defined as `left_edges`; and the topological edges between the right adjacent lanes, defined as `right_edges`. Among them, `predecessor_edge` points to a lane node in the direction of travel of the path, and there may be multiple such nodes; `successor_edge` points to the next lane node in the direction of travel of the path, and there may be multiple such nodes. left_edge: points to the left adjacent lane node; right_edge: Points to the right adjacent lane node.

3. The method according to claim 1, characterized in that, The construction path planning map is built using high-precision maps related to the operation of unmanned vehicles. Lane nodes in Topological relationship between lanes ,include: Determining the probability of lane nodes (lane_nodes); There are two possibilities for a lane node not having left_edges that establish a topological relationship between its left and adjacent lanes: (1) the leftmost lane; (2) lane changes to the left are prohibited. Similarly, if a lane node does not have right_edges (topological relationship edges between its right and adjacent lanes), there are two possibilities: (1) the rightmost lane; (2) the right loop is prohibited. The lane node pointed to by the successor_edge of the topological relationship between the subsequent lanes is called the successor; the lane node pointed to by the left_edge of the topological relationship between the left adjacent lanes is called the left_neighbor; and the lane node pointed to by the right_edge of the topological relationship between the right adjacent lanes is called the right_neighbor. The three are collectively referred to as neighbors. Two-way lanes have both forward and reverse neighbors, while one-way lanes (lane_node) only have forward neighbors.

4. The method according to claim 1, characterized in that, The construction path planning map is built using high-precision maps related to the operation of unmanned vehicles. Lane nodes in Topological relationship between lanes ,include: Definition and cost calculation of topological edges; Each topological edge contains a starting lane node (lane_node), an ending lane node (lane_node), and a toll cost (Call). Call = Cbase + Cadd Cbase is the base cost, which is proportional to the lane length; Cadd is the additional cost, which includes left lane change cost, right lane change cost, left turn cost, right turn cost, and U-turn cost, and can be dynamically adjusted.

5. The method according to claim 1, characterized in that, The method of searching for lane information between the task start and end points based on a high-precision map related to the operation of unmanned vehicles includes: Extract waypoints from the high-precision map related to the task. Equal interval extraction, in which a waypoint is extracted every 0.5 meters. Important task points include: yard bay point, quay crane berth point, lock area point, charging pile point, parking point, or extract a waypoint. Construct a KD-Tree for waypoints and output the data structure of waypoints in k-dimensional Euclidean space; In the KD-Tree structure, find the nearest neighbor waypoint between the task start point and the task end point; Find the lane node corresponding to the nearest neighbor roadpoint (waypoint) between the task start and end points.

6. The method according to claim 1, characterized in that, The aforementioned approach, based on improvements to unmanned vehicles, is specifically designed for port scenarios. The algorithm performs a global path dynamic search, including: Definition and judgment of dynamic cost; Lane changing is prohibited during planning: prohibiting left lane changing sets the additional cost Cadd of the left adjacent edge to infinity (∞); prohibiting right lane changing sets the additional cost Cadd of the right adjacent edge to infinity (∞). During planning, decide whether to pass through the lock zone: If it is decided not to pass through the container loading and unloading lock zone based on the task, set the additional cost Cadd of the successor edge of the lane_node to which the lock zone belongs to infinity ∞.

7. The method according to claim 1, characterized in that, The aforementioned approach, based on improvements to unmanned vehicles, is specifically designed for port scenarios. The algorithm performs a global path dynamic search, including: The search order and planning method for two-way lanes are defined; Positive planning: The algorithm searches for positive neighbors in the current lane_node; positive_neighbors={successors,left_neighbors,right_neighbors} Reverse programming: The algorithm searches from the negative_neighbors of the current lane_node; negative_neighbors={predecessors,left_neighbors,right_neighbors}.

8. A lane-level global path planning device based on a high-precision port map, characterized in that, The device includes: A static high-precision map loading unit is used to acquire static features of the port road environment through a map module and output a static high-precision map of the port environment; wherein, the static features of the road environment include lane lines and road signs; The module adding unit for berth access bridges is used to add templates for berth access bridges at the bow and stern of a ship, respectively, based on a static high-precision port environment map and according to the ship's berthing position. The operation-related high-precision map generation unit is used to overlay static high-precision maps with the template vectors of the upper and lower shore bridges and reconstruct the lane topology to generate operation-related high-precision maps for unmanned vehicles, providing environmental positioning information for subsequent path planning; The path planning graph construction unit is used to construct lane nodes and topological edges in the routing graph using high-precision maps related to autonomous vehicle operations. The lane lookup unit for the start and end points is used to search for lane information between the start and end points of the task based on a high-precision map related to the operation of the unmanned vehicle. A global path dynamic search unit for improving autonomous vehicle-based systems. The algorithm handles two-way lane planning problems and determines whether to prohibit a lane_node from changing lanes or whether to allow passage through container loading / unloading lock zones during planning.

9. A storage medium, characterized in that, The storage medium includes a stored program, wherein, when the program is executed, it controls the device where the storage medium is located to perform the lane-level global path planning method based on a high-precision port map as described in any one of claims 1-7.

10. An electronic device, characterized in that, The device includes at least one processor, and at least one memory and bus connected to the processor; The processor and the memory communicate with each other via the bus. The processor is used to call program instructions in the memory to execute the lane-level global path planning method based on a high-precision port map as described in any one of claims 1-7.

Images