Unmanned vehicle positioning and navigation method based on cooperation of unmanned aerial vehicle and related device thereof

Abstract

The application provides a UAV cooperative unmanned vehicle positioning and navigation method and related equipment thereof, comprising: acquiring three-dimensional point cloud data collected by a UAV in a target area and corresponding real-time dynamic differential positioning information, and generating an original three-dimensional point cloud map with positioning labels; determining coordinate transformation parameters according to the corresponding relationship between the point cloud map coordinates and the world coordinates; compressing the original three-dimensional point cloud map to generate a 2.5-dimensional elevation map with each map grid associated with elevation information and world coordinate labels; in response to a navigation task, generating a target path according to the start point coordinates, the end point coordinates, the world coordinate labels and the elevation information, and converting the path points into world coordinates by using the coordinate transformation parameters to obtain a navigation path; determining lateral deviation and heading deviation according to the real-time dynamic differential positioning information of the unmanned vehicle and the navigation path, and controlling the unmanned vehicle to travel along the navigation path. Through the above method, the coordinate reference of UAV mapping and unmanned vehicle navigation can be unified, and the processing efficiency of map planning can be improved.

Description

Technical Field

[0001] This invention relates to the field of positioning and navigation, and in particular to a method, apparatus, electronic device and storage medium for unmanned vehicle positioning and navigation in collaboration with unmanned aerial vehicles. Background Technology

[0002] With the development of autonomous driving and robotics technologies, unmanned vehicles are increasingly being applied in scenarios such as factory logistics, park inspections, emergency rescue, and field exploration. Autonomous navigation for unmanned vehicles typically relies on environmental maps of the target area, vehicle location information, and path planning results. Unmanned aerial vehicles (UAVs) offer advantages such as wide coverage, fast data collection speed, and minimal impact from ground traffic conditions. They can collect 3D point cloud data of the target area using LiDAR, providing a data foundation for constructing navigation maps for unmanned vehicles.

[0003] Existing autonomous vehicle navigation maps are typically collected by ground-based data collection vehicles navigating within the target area, resulting in lengthy mapping cycles and high map update costs. When using UAVs to collect 3D point cloud data, the point cloud data is usually represented in the UAV's local coordinate system or the point cloud map coordinate system, while the real-time dynamic differential positioning information of autonomous vehicles is usually represented in the world coordinate system. This can easily lead to inconsistencies between the coordinate references of the mapping data, path planning results, and the autonomous vehicle's positioning information. Furthermore, the large volume of 3D point cloud data increases the storage and computational burden on autonomous vehicles when directly used for onboard storage and path planning. Compressing 3D point clouds into elevation maps can easily lose the correspondence between map grids and world coordinates, making paths generated based on elevation maps difficult to use directly for autonomous vehicle positioning and navigation.

[0004] Therefore, how to reduce the storage and computing burden of unmanned vehicle maps while maintaining a unified coordinate benchmark between UAV mapping data, path planning results, and real-time positioning information of unmanned vehicles has become an urgent technical problem to be solved. Summary of the Invention

[0005] This invention provides a drone-assisted unmanned vehicle positioning and navigation method to solve the problem that in existing vehicle navigation schemes, the coordinate reference of drone mapping data and unmanned vehicle real-time dynamic differential positioning information are inconsistent, and the large amount of 3D point cloud data makes it difficult to uniformly transmit and reuse data between mapping, path planning and unmanned vehicle positioning and navigation.

[0006] In a first aspect, embodiments of the present invention provide a method for positioning and navigation of unmanned vehicles in collaboration with unmanned aerial vehicles (UAVs), the method comprising the following steps: Acquire 3D point cloud data collected by the UAV in the target area and the real-time dynamic differential positioning information of the UAV corresponding to the 3D point cloud data, and generate an original 3D point cloud map with positioning tags. Based on the correspondence between the point cloud map coordinates in the original 3D point cloud map and the world coordinates represented by the real-time dynamic differential positioning information of the UAV, the coordinate transformation parameters between the point cloud map coordinate system and the world coordinate system are determined. The original 3D point cloud map is compressed to generate a 2.5D elevation map represented by the point cloud map coordinate system. The world coordinate labels corresponding to each map grid are determined in the 2.5D elevation map based on the coordinate transformation parameters. In response to the navigation task, the 2.5D elevation map is retrieved, and the starting grid and ending grid are determined according to the starting coordinates and ending coordinates corresponding to the navigation task and the world coordinate labels corresponding to each map grid. The target path in the point cloud map coordinate system is generated based on the elevation information corresponding to each map grid. Based on the coordinate transformation parameters, the coordinates of the path points in the target path are converted into world coordinates to obtain the navigation path in the world coordinate system; The system acquires real-time dynamic differential positioning information of the unmanned vehicle during its operation, determines lateral deviation and heading deviation based on the real-time dynamic differential positioning information and the navigation path, and controls the unmanned vehicle to travel along the navigation path based on the lateral deviation and the heading deviation.

[0007] Optionally, the step of acquiring 3D point cloud data collected by the UAV in the target area and the real-time dynamic differential positioning information of the UAV corresponding to the 3D point cloud data, and generating an original 3D point cloud map with positioning tags, includes: The drone is controlled to fly over the target area along a preset route, and multiple point cloud frames are continuously collected by the airborne lidar and the point cloud timestamps corresponding to each point cloud frame are recorded. The airborne real-time dynamic differential positioning module continuously acquires multiple sets of real-time dynamic differential positioning information of the UAV and records the positioning timestamps corresponding to each set of real-time dynamic differential positioning information of the UAV. The real-time dynamic differential positioning information of the UAV includes longitude, latitude, elevation and positioning accuracy factor. Based on the temporal correspondence between each point cloud timestamp and each positioning timestamp, determine the real-time dynamic differential positioning information of the UAV corresponding to each point cloud frame; Each point cloud frame is associated with the corresponding UAV real-time dynamic differential positioning information, and the UAV real-time dynamic differential positioning information is used as the positioning tag of the corresponding point cloud frame to obtain multiple point cloud frames with positioning tags. The original 3D point cloud map is generated based on multiple point cloud frames with location tags.

[0008] Optionally, determining the coordinate transformation parameters between the point cloud map coordinate system and the world coordinate system based on the correspondence between the point cloud map coordinates in the original 3D point cloud map and the world coordinates represented by the real-time dynamic differential positioning information of the UAV includes: Select a first reference point and a second reference point with intervals set from the original 3D point cloud map; Obtain the point cloud map coordinates of the first reference point and the second reference point in the point cloud map coordinate system, respectively; Based on the corresponding real-time dynamic differential positioning information of the UAV, the world coordinates of the first reference point and the second reference point in the world coordinate system are determined. Based on the point cloud map coordinates of the first reference point and the second reference point, determine the first direction vector in the point cloud map coordinate system; Based on the world coordinates of the first reference point and the second reference point, determine the second direction vector in the world coordinate system; Based on the rotation relationship between the first direction vector and the second direction vector, determine the rotation matrix of the point cloud map coordinate system relative to the world coordinate system; Based on the rotation matrix, the point cloud map coordinates of any of the reference points are rotated and transformed, and the translation vector is determined according to the coordinate difference between the rotated point cloud map coordinates and the corresponding world coordinates. The rotation matrix and the translation vector are used as the coordinate transformation parameters.

[0009] Optionally, the step of performing point cloud compression processing on the original 3D point cloud map to generate a 2.5D elevation map represented by the point cloud map coordinate system includes: The original 3D point cloud map is subjected to height threshold filtering to obtain ground and near-ground point clouds; According to the preset grid size, the ground and near-ground point cloud are projected onto the corresponding map grid; The 2.5D elevation map is generated based on the maximum, minimum, and average elevation values ​​of the point cloud points within each map grid.

[0010] Optionally, determining the world coordinate labels corresponding to each map grid in the 2.5D elevation map based on the coordinate transformation parameters includes: Obtain the grid coordinates of each map grid in the point cloud map coordinate system; Based on the coordinate transformation parameters, rotation and translation transformations are performed on the coordinates of each grid to obtain the global coordinates of each map grid in the world coordinate system. The longitude, latitude, and elevation corresponding to each of the global coordinates are used as the world coordinate labels for the corresponding map grid.

[0011] Optionally, in response to the navigation task, retrieving the 2.5D elevation map and determining the starting and ending grids based on the starting and ending coordinates corresponding to the navigation task and the world coordinate labels corresponding to each map grid includes: In response to the navigation task, the starting coordinates and ending coordinates corresponding to the navigation task are obtained, and the starting coordinates and ending coordinates are both located in the world coordinate system; Retrieve the 2.5D elevation map corresponding to the target area and read the world coordinate labels corresponding to each map grid. The starting point coordinates are matched with each of the world coordinate labels, and the map grid corresponding to the world coordinate label that matches the starting point coordinates is determined as the starting point grid. The endpoint coordinates are matched with each of the world coordinate labels, and the map grid corresponding to the world coordinate label that matches the endpoint coordinates is determined as the endpoint grid.

[0012] Optionally, the step of converting the coordinates of path points in the target path into world coordinates based on the coordinate transformation parameters to obtain the navigation path in the world coordinate system includes: Extract multiple path points and the path point coordinates of each path point in the point cloud map coordinate system sequentially from the target path; Obtain the rotation matrix and translation vector of the point cloud map coordinate system relative to the world coordinate system from the coordinate transformation parameters; Based on the rotation matrix, the coordinates of each path point are rotated and transformed, and based on the translation vector, the coordinates of the rotated path points are translated and transformed to obtain the world coordinates corresponding to each path point. The navigation path is obtained by combining the world coordinates according to the order in which the path points are arranged in the target path.

[0013] Secondly, embodiments of the present invention also provide a drone-assisted unmanned vehicle positioning and navigation device, the drone-assisted unmanned vehicle positioning and navigation device comprising: The acquisition module is used to acquire three-dimensional point cloud data collected by the UAV in the target area and the real-time dynamic differential positioning information of the UAV corresponding to the three-dimensional point cloud data, and generate an original three-dimensional point cloud map with positioning tags. The determination module is used to determine the coordinate transformation parameters between the point cloud map coordinate system and the world coordinate system based on the correspondence between the point cloud map coordinates in the original 3D point cloud map and the world coordinates represented by the real-time dynamic differential positioning information of the UAV. The generation module is used to perform point cloud compression processing on the original 3D point cloud map to generate a 2.5D elevation map represented by the point cloud map coordinate system, and to determine the world coordinate labels corresponding to each map grid in the 2.5D elevation map based on the coordinate transformation parameters. The response module is used to respond to the navigation task, retrieve the 2.5D elevation map, determine the starting grid and ending grid according to the starting coordinates and ending coordinates corresponding to the navigation task and the world coordinate labels corresponding to each map grid, and generate the target path in the point cloud map coordinate system based on the elevation information corresponding to each map grid. The transformation module is used to convert the coordinates of path points in the target path into world coordinates based on the coordinate transformation parameters, so as to obtain the navigation path in the world coordinate system; The navigation module is used to acquire the onboard real-time dynamic differential positioning information of the unmanned vehicle during its driving process, determine the lateral deviation and heading deviation based on the onboard real-time dynamic differential positioning information and the navigation path, and control the unmanned vehicle to drive along the navigation path based on the lateral deviation and the heading deviation.

[0014] Thirdly, embodiments of the present invention provide an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps in the drone-cooperative unmanned vehicle positioning and navigation method provided in embodiments of the present invention.

[0015] Fourthly, embodiments of the present invention provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps in the unmanned vehicle positioning and navigation method with UAV collaboration provided in the embodiments of the present invention.

[0016] In this embodiment of the invention, 3D point cloud data collected by the UAV in the target area and corresponding real-time dynamic differential positioning information are acquired to generate an original 3D point cloud map with positioning tags. Coordinate transformation parameters are determined based on the correspondence between point cloud map coordinates and world coordinates. The original 3D point cloud map is compressed to generate a 2.5D elevation map with associated elevation information and world coordinate tags for each map grid. Responding to the navigation task, a target path is generated based on the starting point coordinates, ending point coordinates, world coordinate tags, and elevation information. The path points are converted to world coordinates using the coordinate transformation parameters to obtain the navigation path. Lateral and heading deviations are determined based on the UAV's real-time dynamic differential positioning information and the navigation path, controlling the UAV to travel along the navigation path. This method unifies the coordinate references for UAV mapping and UAV navigation, improving the processing efficiency of map planning. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart of a drone-assisted unmanned vehicle positioning and navigation method provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of another drone-assisted unmanned vehicle positioning and navigation device provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0019] 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. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] like Figure 1 As shown, Figure 1 This is a flowchart of a drone-assisted unmanned vehicle positioning and navigation method provided by an embodiment of the present invention. The drone-assisted unmanned vehicle positioning and navigation method includes the following steps: 101. Acquire the 3D point cloud data collected by the UAV in the target area and the real-time dynamic differential positioning information of the UAV corresponding to the 3D point cloud data, and generate an original 3D point cloud map with positioning tags.

[0021] In this embodiment of the invention, the above-mentioned UAV-coordinated unmanned vehicle positioning and navigation method can be applied to a UAV-coordinated unmanned vehicle positioning and navigation platform. The UAV-coordinated unmanned vehicle positioning and navigation platform can have functions such as UAV-coordinated unmanned vehicle positioning and navigation data processing, UAV-coordinated unmanned vehicle positioning and navigation data transmission and reception, and UAV-coordinated unmanned vehicle positioning and navigation data memory storage. It can be built based on a server or server cluster. The server or server cluster can be an electronic device with UAV-coordinated unmanned vehicle positioning and navigation data processing capabilities.

[0022] The target area mentioned above can be the spatial area where the unmanned vehicle needs to perform positioning and navigation tasks, such as industrial parks, logistics parks, inspection areas, or field operation areas.

[0023] The aforementioned 3D point cloud data can be understood as a set of spatial points formed after the target area is scanned by a lidar mounted on a drone. Each point cloud point includes its 3D coordinates in the lidar coordinate system or the point cloud map coordinate system, which are used to characterize the spatial location of the ground, roads, buildings and other environmental objects within the target area.

[0024] The aforementioned real-time dynamic differential positioning information for UAVs refers to the positioning data obtained by the UAV through its onboard real-time dynamic differential positioning module during the acquisition of 3D point cloud data. This data includes longitude, latitude, elevation, and positioning accuracy factor. The positioning accuracy factor is used to characterize the positioning quality of the corresponding positioning information.

[0025] The aforementioned positioning tags can be understood as positioning information markers associated with point cloud frames. They are used to record the world coordinates of the UAV when the point cloud frame is acquired, enabling the point cloud frame to establish a correspondence with the corresponding UAV real-time dynamic differential positioning information.

[0026] The aforementioned original 3D point cloud map refers to a point cloud map formed by summarizing multiple point cloud frames and their positioning tags collected by the UAV, without undergoing coordinate transformation and point cloud compression.

[0027] In one possible embodiment, the aforementioned drone-assisted unmanned vehicle positioning and navigation platform can send flight control data to the drone, enabling it to fly along a preset route over the target area. Taking an industrial park as an example, the preset route can be a zigzag pattern, allowing the drone to sequentially scan the roads and surrounding environment within the industrial park. The drone's flight altitude can be set to 80 meters above the ground, the forward overlap rate can be set to 70%, and the lateral overlap rate can be set to 50% to reduce the occurrence of missing point clouds between adjacent scanned areas.

[0028] Specifically, the aforementioned drone-assisted unmanned vehicle positioning and navigation platform can control the drone to continuously scan the target area using an airborne LiDAR. The airborne LiDAR outputs multiple point cloud frames according to a preset point cloud acquisition frequency and records a corresponding point cloud timestamp for each point cloud frame. The point cloud timestamp is used to indicate the acquisition time of the corresponding point cloud frame.

[0029] While the UAVs collect point cloud frames, the aforementioned UAV-assisted unmanned vehicle positioning and navigation platform can continuously acquire real-time dynamic differential positioning information of the UAVs through an onboard real-time dynamic differential positioning module, and record the corresponding positioning timestamps for each group of UAVs' real-time dynamic differential positioning information. For example, the onboard lidar can output point cloud frames at a frequency of 10 Hz, and the onboard real-time dynamic differential positioning module can output real-time dynamic differential positioning information of the UAVs at a frequency of 20 Hz.

[0030] After data collection is completed, the aforementioned drone-assisted unmanned vehicle positioning and navigation platform can time-align multiple point cloud frames with multiple sets of real-time dynamic differential positioning information of drones based on the temporal correspondence between each point cloud timestamp and each positioning timestamp, thereby determining the longitude, latitude, elevation and positioning accuracy factor corresponding to each point cloud frame at the time of collection.

[0031] The aforementioned drone-assisted unmanned vehicle positioning and navigation platform can use time-aligned real-time dynamic differential positioning information of the drone as positioning tags for corresponding point cloud frames. It then aggregates multiple point cloud frames with positioning tags to generate an original 3D point cloud map. This original 3D point cloud map simultaneously retains the 3D coordinates of each point, the point cloud timestamp of each point cloud frame, and the corresponding positioning tag, providing a data foundation for subsequently determining the coordinate transformation parameters between the point cloud map coordinate system and the world coordinate system.

[0032] Through the above methods and steps, point cloud data of the target area can be quickly collected using UAVs, and each point cloud frame can be accurately matched with the real-time dynamic differential positioning information of the UAV at the time of collection. This provides raw map data with a world coordinate correlation for subsequent point cloud coordinate transformation, 2.5D elevation map generation, and unmanned vehicle positioning and navigation.

[0033] 102. Based on the correspondence between the point cloud map coordinates in the original 3D point cloud map and the world coordinates represented by the real-time dynamic differential positioning information of the UAV, determine the coordinate transformation parameters between the point cloud map coordinate system and the world coordinate system.

[0034] In this embodiment of the invention, the above-mentioned point cloud map coordinates can be understood as the position coordinates of point cloud points or reference points in the point cloud map coordinate system, which uses the map origin determined during the UAV mapping process as the coordinate origin.

[0035] The aforementioned world coordinates can be longitude, latitude, and elevation in the world coordinate system, used to characterize the actual geographical location of the corresponding point cloud point or reference point in the target area.

[0036] The above correspondence refers to the pairing relationship between the point cloud map coordinates of the same reference point in the point cloud map coordinate system and its world coordinates in the world coordinate system.

[0037] The coordinate transformation parameters described above are used to convert point cloud map coordinates to world coordinates, including rotation matrices and translation vectors. The rotation matrix represents the directional difference between the two coordinate systems, and the translation vector represents the positional difference between the origins of the two coordinate systems.

[0038] In one possible embodiment, the aforementioned UAV-assisted unmanned vehicle positioning and navigation platform can select a first reference point and a second reference point set at intervals from the original 3D point cloud map. Taking an industrial park as an example, the aforementioned UAV-assisted unmanned vehicle positioning and navigation platform can select two reference points located at the two ends of the park's longer diagonal to increase the distance between the two reference points and reduce the impact of small positional errors on the direction relationship calculation results.

[0039] Specifically, the aforementioned drone-assisted unmanned vehicle positioning and navigation platform can obtain the point cloud map coordinates of the first and second reference points in the point cloud map coordinate system, and read the real-time dynamic differential positioning information of the drones for the first and second reference points from the corresponding positioning tags to determine the world coordinates of the two reference points in the world coordinate system. Thus, the point cloud map coordinates and world coordinates of the same reference point constitute a set of corresponding coordinates.

[0040] The aforementioned drone-assisted unmanned vehicle positioning and navigation platform can calculate a first direction vector from the first reference point to the second reference point based on the point cloud map coordinates of the first and second reference points; and calculate a second direction vector from the first reference point to the second reference point based on the world coordinates of the first and second reference points.

[0041] The aforementioned UAV-assisted unmanned vehicle positioning and navigation platform can also determine a rotation matrix that can rotate the first direction vector to the second direction vector based on the directional difference between the first and second direction vectors. Subsequently, the UAV-assisted unmanned vehicle positioning and navigation platform can use the rotation matrix to perform a rotation transformation on the point cloud map coordinates of any reference point, and calculate the coordinate difference between the rotated point cloud map coordinates and the corresponding world coordinates of the reference point, and determine the coordinate difference as a translation vector.

[0042] The aforementioned drone-assisted unmanned vehicle positioning and navigation platform can determine the rotation matrix and translation vector as coordinate transformation parameters between the point cloud map coordinate system and the world coordinate system, and store the coordinate transformation parameters in association with the original 3D point cloud map for subsequent use when determining the world coordinate labels of the map grid and transforming the coordinates of path points in the target path.

[0043] By using the above methods and steps, the transformation relationship between the point cloud map coordinate system and the world coordinate system can be established based on the corresponding coordinates of the same reference point in two coordinate systems. This allows the subsequently generated map grids and navigation paths to use a unified world coordinate reference, reducing positional deviations caused by inconsistencies in coordinate references between mapping, path generation, and autonomous vehicle positioning and navigation.

[0044] 103. Perform point cloud compression processing on the original 3D point cloud map to generate a 2.5D elevation map represented by the point cloud map coordinate system, and determine the world coordinate label corresponding to each map grid in the 2.5D elevation map based on the coordinate transformation parameters.

[0045] In this embodiment of the invention, the above-mentioned drone-assisted unmanned vehicle positioning and navigation platform can filter and rasterize the point cloud points in the original three-dimensional point cloud map according to their spatial location, and use the elevation information of the point cloud points in each grid to represent the corresponding area, thereby reducing the amount of point cloud data that needs to be stored and processed.

[0046] The aforementioned 2.5D elevation map refers to a map that uses two-dimensional map grids to represent the planar location of the target area, and records the elevation information of the corresponding area in each map grid. The aforementioned 2.5D elevation map still uses a point cloud map coordinate system to represent the planar location of each map grid.

[0047] The aforementioned map grid refers to a map unit formed by dividing the target area according to a preset grid size. Each map grid corresponds to a planar range within the target area and records the elevation information formed by point cloud points falling within that planar range.

[0048] The aforementioned world coordinate labels refer to the world coordinate information associated with the map raster, including longitude, latitude, and elevation, which are used to establish the correspondence between the map raster and the actual geographical location of the target area.

[0049] In one possible embodiment, the aforementioned UAV-assisted unmanned vehicle positioning and navigation platform can statistically analyze the elevation distribution of each point cloud point in the original 3D point cloud map and extract ground point clouds from the original 3D point cloud map according to a preset percentage threshold. For example, the aforementioned UAV-assisted unmanned vehicle positioning and navigation platform can identify point cloud points with elevation values ​​within the lowest 8% range as ground point clouds and determine the ground elevation reference surface based on the ground point clouds.

[0050] Specifically, the aforementioned drone-assisted unmanned vehicle positioning and navigation platform can perform height threshold filtering on the original 3D point cloud map based on the ground elevation datum, retaining ground and near-ground point clouds to reduce the impact of high-altitude point clouds unrelated to the unmanned vehicle's passage area on the elevation map generation process.

[0051] The aforementioned drone-assisted unmanned vehicle positioning and navigation platform can divide the horizontal plane of the point cloud map coordinate system according to a preset grid size, and project the ground and near-ground point clouds onto the corresponding map grid based on the planar coordinates of each point cloud point in the point cloud map coordinate system. For example, the preset grid size can be set to 0.2 meters × 0.2 meters.

[0052] The aforementioned UAV-assisted unmanned vehicle positioning and navigation platform can statistically analyze the point cloud points falling within each map grid, and determine the maximum, minimum, and average elevation values ​​of the point cloud points within each map grid as the elevation information of the corresponding map grid. For map grids where no point cloud points fall, the aforementioned UAV-assisted unmanned vehicle positioning and navigation platform can interpolate and fill the grid based on the elevation information of adjacent map grids, thereby generating a continuous 2.5D elevation map.

[0053] After determining the elevation information of each map grid, the aforementioned UAV-assisted unmanned vehicle positioning and navigation platform can obtain the grid coordinates of each map grid in the point cloud map coordinate system, and use the rotation matrix and translation vector in the coordinate transformation parameters to perform rotation and translation transformations on the grid coordinates to obtain the world coordinates corresponding to each map grid.

[0054] The aforementioned drone-assisted unmanned vehicle positioning and navigation platform can use the longitude, latitude, and elevation corresponding to each world coordinate as the world coordinate labels of the corresponding map grid, and associate and store each world coordinate label with the grid coordinates and elevation information of the corresponding map grid to obtain a 2.5-dimensional elevation map that simultaneously contains point cloud map coordinates, elevation information, and world coordinate labels.

[0055] Through the above methods and steps, the original 3D point cloud map with a large amount of data can be compressed into a 2.5D elevation map represented by map rasters, which reduces the amount of map data while retaining the elevation information of the target area. Furthermore, each map raster establishes a correspondence with the actual geographical location through world coordinate labels, thereby providing a data basis for determining the starting and ending raster based on the starting and ending coordinates of the navigation task.

[0056] 104. In response to the navigation task, retrieve the 2.5D elevation map, determine the starting grid and ending grid according to the starting coordinates and ending coordinates corresponding to the navigation task and the world coordinate labels corresponding to each map grid, and generate the target path in the point cloud map coordinate system based on the elevation information corresponding to each map grid.

[0057] In this embodiment of the invention, the navigation task can be understood as task data used to instruct the unmanned vehicle to travel from the starting position to the target position, including at least the target area, the starting coordinates, and the ending coordinates.

[0058] The aforementioned starting point coordinates and ending point coordinates are used to represent the position when the autonomous vehicle begins to perform the navigation task and the position it needs to reach, respectively, and are both represented using coordinates in the world coordinate system.

[0059] The aforementioned starting point grid and ending point grid are map grids in the 2.5D elevation map that correspond to the starting point coordinates and ending point coordinates, respectively. They are used to convert world coordinates in the navigation task into grid locations that can be used for path generation in the 2.5D elevation map.

[0060] The aforementioned target path can be understood as path data composed of multiple path points from the starting grid to the ending grid in the order of travel, with each path point represented by coordinates in the point cloud map coordinate system.

[0061] In one possible embodiment, the aforementioned drone-assisted unmanned vehicle positioning and navigation platform can, after receiving a navigation task, extract the target area, starting coordinates, and ending coordinates from the navigation task, and retrieve the corresponding 2.5D elevation map from the storage medium based on the target area.

[0062] Specifically, the aforementioned drone-assisted unmanned vehicle positioning and navigation platform can read the world coordinate labels corresponding to each map grid and perform coordinate matching between the start-point coordinates and the end-point coordinates and each world coordinate label. In one possible embodiment, the drone-assisted unmanned vehicle positioning and navigation platform can calculate the coordinate distances between the start-point coordinates and the end-point coordinates and each world coordinate label, and determine the map grid corresponding to the world coordinate label with the smallest coordinate distance to the start-point coordinates as the start-point grid, and determine the map grid corresponding to the world coordinate label with the smallest coordinate distance to the end-point coordinates as the end-point grid.

[0063] After determining the starting and ending grids, the aforementioned UAV-assisted unmanned vehicle positioning and navigation platform can determine the elevation change rate between adjacent map grids based on the elevation information and grid spacing. It can then determine whether a corresponding map grid is passable by the unmanned vehicle based on the elevation change rate. For example, map grids with a slope of less than 15 degrees corresponding to the elevation change rate can be identified as passable grids, while those that do not meet this passability condition can be identified as impassable grids.

[0064] The aforementioned drone-assisted unmanned vehicle positioning and navigation platform can use the starting grid as the initial position for path search and the ending grid as the target position. It employs the A* algorithm to search for paths within each passable grid, obtaining an initial path from the starting grid to the ending grid. Each path point in the initial path is represented using the grid coordinates of its corresponding map grid in the point cloud map coordinate system.

[0065] To reduce the impact of excessive directional changes between adjacent path points on the autonomous vehicle's driving process, the aforementioned UAV-assisted autonomous vehicle positioning and navigation platform can extract multiple key path points as control points from the initial path and use a third-order B-spline interpolation method to smooth the path of each control point, obtaining the target path in the point cloud map coordinate system. For example, the smoothing weight coefficient can be set to 0.6, and multiple ordered path points can be generated according to the preset path point intervals based on the smoothed curve.

[0066] Taking the navigation task in an industrial park as an example, when an unmanned vehicle needs to travel from point A to point B in the park, the aforementioned drone-assisted unmanned vehicle positioning and navigation platform can determine the starting grid and the ending grid according to the world coordinates of point A and point B, respectively, and generate an initial path based on the passable grid in the 2.5D elevation map. After path smoothing, a target path composed of multiple path points is obtained.

[0067] By using the above methods and steps, the starting and ending coordinates of the navigation task can be accurately mapped to a 2.5-dimensional elevation map using the world coordinate labels of the map grid. Combined with the elevation information of the map grid, areas that do not meet the passage conditions can be avoided. Thus, a smooth target path suitable for autonomous vehicle driving can be generated by reusing the stored 2.5-dimensional elevation map.

[0068] 105. Based on the coordinate transformation parameters, convert the coordinates of the path points in the target path to world coordinates to obtain the navigation path in the world coordinate system.

[0069] In this embodiment of the invention, the above-mentioned drone-assisted unmanned vehicle positioning and navigation platform can extract the coordinates of each path point in the point cloud map coordinate system according to the order of each path point in the target path, and retrieve the coordinate transformation parameters corresponding to the current 2.5-dimensional elevation map.

[0070] Specifically, the aforementioned drone-assisted unmanned vehicle positioning and navigation platform can use the rotation matrix in the coordinate transformation parameters to perform rotation transformation on the coordinates of each path point, and then use the translation vector to perform translation transformation on the rotated path point coordinates to obtain the world coordinates of each path point in the world coordinate system. The world coordinates of each path point include the corresponding longitude, latitude, and elevation.

[0071] The aforementioned drone-assisted unmanned vehicle positioning and navigation platform can combine multiple world coordinates obtained by transformation according to the original arrangement order of each waypoint in the target path to obtain a navigation path in the world coordinate system, and load the navigation path into the positioning and navigation module of the unmanned vehicle.

[0072] For example, if the target path includes 280 path points arranged in the order of autonomous vehicle travel, the aforementioned UAV-assisted autonomous vehicle positioning and navigation platform can perform rotation and translation transformations on the coordinates of the 280 path points to obtain the corresponding 280 world coordinates, and generate a navigation path according to the original arrangement.

[0073] The above methods and steps can convert the target path in the point cloud map coordinate system into a navigation path that uses the same coordinate reference as the vehicle's real-time dynamic differential positioning information, thus providing a coordinate basis for subsequently determining the lateral and heading deviations of the unmanned vehicle relative to the navigation path.

[0074] 106. Obtain the onboard real-time dynamic differential positioning information of the unmanned vehicle during its operation, determine the lateral deviation and heading deviation based on the onboard real-time dynamic differential positioning information and the navigation path, and control the unmanned vehicle to travel along the navigation path based on the lateral deviation and heading deviation.

[0075] In this embodiment of the invention, the above-mentioned vehicle-mounted real-time dynamic differential positioning information is used to characterize the current position and current heading of the unmanned vehicle during its driving process. The current position is represented by coordinates in the world coordinate system.

[0076] The aforementioned lateral deviation can be understood as the lateral distance of the autonomous vehicle's current position relative to the navigation path, used to characterize the degree to which the autonomous vehicle deviates from one side of the navigation path.

[0077] The aforementioned heading deviation can be understood as the angle between the current heading of the autonomous vehicle and the path direction of the navigation path at the corresponding position, used to characterize the degree of deviation of the autonomous vehicle's driving direction from the navigation path direction.

[0078] In one possible embodiment, the aforementioned drone-assisted unmanned vehicle positioning and navigation platform can continuously acquire onboard real-time dynamic differential positioning information at a preset positioning update frequency through the onboard real-time dynamic differential positioning module of the unmanned vehicle. For example, the onboard real-time dynamic differential positioning module can update the current position and current heading of the unmanned vehicle at a frequency of 10 Hz.

[0079] Specifically, the aforementioned drone-assisted unmanned vehicle positioning and navigation platform can determine the target path point closest to the current location from the navigation path based on the current location of the unmanned vehicle, and determine the path tangent direction corresponding to the location of the target path point.

[0080] The aforementioned drone-assisted unmanned vehicle positioning and navigation platform can determine the lateral deviation based on the vertical distance between the unmanned vehicle's current position and the target path point, and determine the heading deviation based on the angle between the unmanned vehicle's current heading and the path tangent direction.

[0081] In one possible embodiment, the aforementioned UAV-assisted unmanned vehicle positioning and navigation platform can input lateral deviation and heading deviation into a model predictive controller. The model predictive controller then predicts the unmanned vehicle's driving state at multiple subsequent moments and generates throttle control and steering control commands based on the prediction results. For example, the prediction time domain of the model predictive controller can be set to 20 steps, and the control time domain can be set to 10 steps. In other embodiments, corresponding vehicle control commands can also be generated based on the lateral deviation and heading deviation using proportional-integral-derivative (PID) control.

[0082] The aforementioned drone-assisted unmanned vehicle positioning and navigation platform can send throttle control commands to the unmanned vehicle's drive mechanism and steering control commands to its steering mechanism to adjust the vehicle's speed and direction. During operation, the platform continuously updates the vehicle's real-time dynamic differential positioning information and re-determines lateral and heading deviations until the vehicle reaches the destination point on the navigation path.

[0083] For example, during the transportation of materials in an industrial park, the aforementioned drone-assisted unmanned vehicle positioning and navigation platform can use model predictive control to keep the lateral deviation of the unmanned vehicle relative to the navigation path within ±0.1 meters, enabling the unmanned vehicle to travel from the starting point of the park to the end point along the navigation path.

[0084] By using the above methods and steps, the real-time position of the unmanned vehicle can be continuously compared with the navigation path in a unified world coordinate system, and the driving status of the unmanned vehicle can be adjusted in a timely manner according to the lateral deviation and heading deviation, thereby improving the accuracy and stability of the unmanned vehicle driving along the navigation path.

[0085] In this embodiment of the invention, 3D point cloud data collected by the UAV in the target area and corresponding real-time dynamic differential positioning information are acquired to generate an original 3D point cloud map with positioning tags. Coordinate transformation parameters are determined based on the correspondence between point cloud map coordinates and world coordinates. The original 3D point cloud map is compressed to generate a 2.5D elevation map with associated elevation information and world coordinate tags for each map grid. Responding to the navigation task, a target path is generated based on the starting point coordinates, ending point coordinates, world coordinate tags, and elevation information. The path points are converted to world coordinates using the coordinate transformation parameters to obtain the navigation path. Lateral and heading deviations are determined based on the UAV's real-time dynamic differential positioning information and the navigation path, controlling the UAV to travel along the navigation path. This method unifies the coordinate references for UAV mapping and UAV navigation, improving the processing efficiency of map planning.

[0086] Optionally, in the steps of acquiring 3D point cloud data collected by the UAV in the target area and the corresponding real-time dynamic differential positioning information of the UAV, and generating an original 3D point cloud map with positioning tags, the UAV can be controlled to fly along a preset route over the target area, and multiple point cloud frames can be continuously acquired by the airborne LiDAR and the point cloud timestamps corresponding to each point cloud frame can be recorded; multiple sets of real-time dynamic differential positioning information of the UAV can be continuously acquired by the airborne real-time dynamic differential positioning module and the positioning timestamps corresponding to each set of real-time dynamic differential positioning information of the UAV can be recorded. The real-time dynamic differential positioning information of the UAV includes longitude, latitude, elevation and positioning accuracy factor; the real-time dynamic differential positioning information of the UAV corresponding to each point cloud frame can be determined according to the temporal correspondence between each point cloud timestamp and each positioning timestamp; each point cloud frame is associated with the corresponding real-time dynamic differential positioning information of the UAV, and the real-time dynamic differential positioning information of the UAV is used as the positioning tag of the corresponding point cloud frame to obtain multiple point cloud frames with positioning tags; and an original 3D point cloud map is generated based on multiple point cloud frames with positioning tags.

[0087] In this embodiment of the invention, the above-mentioned positioning accuracy factor is used to characterize the influence of satellite spatial distribution on positioning accuracy, and the position accuracy factor (PDOP value) can be used.

[0088] In one possible embodiment, the aforementioned drone-assisted unmanned vehicle positioning and navigation platform can be applied to large industrial parks. The drones are controlled to fly at an altitude of 80 meters above the ground in a "bow"-shaped flight path, with a 70% forward overlap and a 50% lateral overlap. During flight, an airborne lidar collects point cloud frames at a frequency of 10Hz, and an airborne real-time dynamic differential positioning module collects longitude, latitude, altitude, and PDOP values ​​at a frequency of 20Hz.

[0089] After the flight is completed, the aforementioned UAV-assisted unmanned vehicle positioning and navigation platform can use the point cloud timestamps of each point cloud frame as time indexes to determine the corresponding UAV real-time dynamic differential positioning information from the positioning information arranged by positioning timestamps, and associate it with the corresponding point cloud frame. Subsequently, multiple point cloud frames with positioning tags are collected in the order of acquisition to generate an original 3D point cloud map containing approximately 230 million point cloud points.

[0090] This provides each point cloud frame with traceable world coordinate positioning data, offering a data foundation for subsequently determining the coordinate transformation parameters between the point cloud map coordinate system and the world coordinate system.

[0091] Optionally, the step of determining the coordinate transformation parameters between the point cloud map coordinate system and the world coordinate system based on the correspondence between the point cloud map coordinates in the original 3D point cloud map and the world coordinates represented by the real-time dynamic differential positioning information of the UAV, further includes: selecting a first reference point and a second reference point set at intervals from the original 3D point cloud map; obtaining the point cloud map coordinates of the first reference point and the second reference point in the point cloud map coordinate system respectively; determining the world coordinates of the first reference point and the second reference point in the world coordinate system based on the corresponding real-time dynamic differential positioning information of the UAV; determining a first direction vector in the point cloud map coordinate system based on the point cloud map coordinates of the first reference point and the second reference point; determining a second direction vector in the world coordinate system based on the world coordinates of the first reference point and the second reference point; determining the rotation matrix of the point cloud map coordinate system relative to the world coordinate system based on the rotation relationship between the first direction vector and the second direction vector; performing a rotation transformation on the point cloud map coordinates of any reference point based on the rotation matrix, and determining a translation vector based on the coordinate difference between the rotated point cloud map coordinates and the corresponding world coordinates; and using the rotation matrix and the translation vector as coordinate transformation parameters.

[0092] In this embodiment of the invention, the first direction vector is used to represent the direction of coordinate change of the point cloud map from the first reference point to the second reference point; the second direction vector is used to represent the direction of coordinate change of the same two reference points in the world coordinate system. The coordinate difference can be understood as the difference between the corresponding world coordinates and the point cloud map coordinates after rotation transformation in each coordinate component, used to characterize the origin offset between the two coordinate systems.

[0093] In one possible embodiment, the drone-assisted unmanned vehicle positioning and navigation platform can select two points located at the two ends of the longest diagonal of the industrial park from the original three-dimensional point cloud map corresponding to the industrial park, respectively, so that the two reference points are set at intervals.

[0094] The aforementioned drone-assisted unmanned vehicle positioning and navigation platform can read the coordinates of two reference points in the aforementioned point cloud map coordinate system, and obtain the longitude, latitude and elevation of the two reference points in the aforementioned world coordinate system based on the real-time dynamic differential positioning information of the drones associated with the two reference points, thereby forming a coordinate correspondence relationship of the same reference point in the two coordinate systems.

[0095] Specifically, the aforementioned UAV-assisted unmanned vehicle positioning and navigation platform can determine the first direction vector pointing from the first reference point to the second reference point based on the point cloud map coordinates of the two reference points, and determine the second direction vector based on the world coordinates of the two reference points. A rotation matrix is ​​determined by rotating the first direction vector to align with the second direction vector. Subsequently, the point cloud map coordinates of the first or second reference point are rotated using the rotation matrix, and the world coordinates corresponding to the reference point are compared with the rotated point cloud map coordinates. The resulting coordinate difference is determined as the translation vector. Finally, the rotation matrix and the translation vector are associated with and saved with the original 3D point cloud map as coordinate transformation parameters used for subsequent map raster world coordinate annotation and path point coordinate transformation.

[0096] Through the above processing, the directional transformation relationship and position offset relationship of the point cloud map coordinate system relative to the world coordinate system can be uniformly determined.

[0097] Optionally, the step of performing point cloud compression processing on the original 3D point cloud map to generate a 2.5D elevation map represented by the point cloud map coordinate system further includes performing height threshold filtering on the original 3D point cloud map to obtain ground and near-ground point clouds; projecting the ground and near-ground point clouds onto the corresponding map grids according to the preset grid size; and generating a 2.5D elevation map based on the maximum, minimum, and average elevation values ​​of the point cloud points in each map grid.

[0098] In this embodiment of the invention, the above-mentioned height threshold filtering can be understood as using the ground elevation reference surface as a comparison reference, comparing the elevation value of each point cloud point with the above-mentioned ground elevation reference surface, and retaining the ground and near-ground point clouds that meet the preset height range.

[0099] In one possible embodiment, the aforementioned UAV-assisted unmanned vehicle positioning and navigation platform can statistically analyze the elevation distribution of all point cloud points in the original 3D point cloud map, extract the 8% of point cloud points with the lowest elevation values ​​as ground point clouds, and determine the ground elevation datum based on the ground point clouds. Subsequently, the UAV-assisted unmanned vehicle positioning and navigation platform can perform height threshold filtering on the point cloud points based on the ground elevation datum to obtain the aforementioned ground and near-ground point clouds used to characterize the road and near-ground environment.

[0100] Specifically, the aforementioned drone-assisted unmanned vehicle positioning and navigation platform can set the preset grid size to 0.2 meters × 0.2 meters, and project the ground and near-ground point clouds onto the corresponding map grids according to the planar coordinates of each point cloud point in the aforementioned point cloud map coordinate system.

[0101] For each of the aforementioned map grids, the UAV-assisted unmanned vehicle positioning and navigation platform can count the point cloud points falling within that grid, determine their maximum, minimum, and average elevation values, and associate these three types of elevation values ​​with the corresponding map grids to form the aforementioned 2.5D elevation map. For map grids without any point cloud points falling within them, bilinear interpolation can be performed based on the elevation values ​​of adjacent map grids to supplement the corresponding elevation information.

[0102] Through the above processing, the original 3D point cloud data can be converted into elevation data organized by map grid, which reduces the amount of data required for subsequent storage and path planning while preserving ground undulation information.

[0103] Optionally, the step of determining the world coordinate label corresponding to each map grid in the 2.5D elevation map based on coordinate transformation parameters further includes obtaining the grid coordinates of each map grid in the point cloud map coordinate system; performing rotation and translation transformations on each grid coordinate based on the coordinate transformation parameters to obtain the global coordinates of each map grid in the world coordinate system; and using the longitude, latitude, and elevation corresponding to each global coordinate as the world coordinate label of the corresponding map grid.

[0104] In this embodiment of the invention, the grid coordinates are used to characterize the position of the corresponding map grid in the point cloud map coordinate system, and can be composed of the planar coordinates of the map grid and the corresponding elevation information. The planar coordinates can be determined based on the map origin, the preset grid size, and the row and column positions of the corresponding map grid; the elevation information can be the average elevation value recorded by the corresponding map grid.

[0105] In one possible embodiment, the aforementioned UAV-assisted unmanned vehicle positioning and navigation platform can sequentially read each map grid in the aforementioned 2.5D elevation map and determine the grid coordinates of each map grid in the aforementioned point cloud map coordinate system. Subsequently, the aforementioned UAV-assisted unmanned vehicle positioning and navigation platform can retrieve the aforementioned rotation matrix and translation vector associated with and stored in the aforementioned 2.5D elevation map, first perform rotation transformation on the grid coordinates based on the aforementioned rotation matrix, and then perform position translation on the rotated grid coordinates based on the aforementioned translation vector to obtain the global coordinates of each map grid in the aforementioned world coordinate system.

[0106] The aforementioned drone-assisted unmanned vehicle positioning and navigation platform can extract longitude, latitude, and elevation from the aforementioned global coordinates, and use these longitude, latitude, and elevation as the world coordinate labels of the corresponding map grids, and save them in association with the corresponding map grids.

[0107] Therefore, each map grid in the aforementioned 2.5D elevation map can be traced back to world coordinate information, providing a unified coordinate basis for matching the corresponding map grid based on the starting and ending coordinates in the navigation task.

[0108] Optionally, in response to a navigation task, the steps of retrieving a 2.5D elevation map and determining the starting and ending grids based on the starting and ending coordinates corresponding to the navigation task and the world coordinate labels corresponding to each map grid further include: responding to a navigation task and obtaining the starting and ending coordinates corresponding to the navigation task, where both the starting and ending coordinates are located in the world coordinate system; retrieving the 2.5D elevation map corresponding to the target area and reading the world coordinate labels corresponding to each map grid; performing coordinate matching between the starting coordinates and each world coordinate label, and determining the map grid corresponding to the world coordinate label that matches the starting coordinates as the starting grid; performing coordinate matching between the ending coordinates and each world coordinate label, and determining the map grid corresponding to the world coordinate label that matches the ending coordinates as the ending grid.

[0109] In this embodiment of the invention, the above coordinate matching can be understood as: comparing the longitude, latitude and elevation in the starting point coordinates or ending point coordinates with each of the above world coordinate labels, determining the world coordinate labels with the same coordinates or the smallest coordinate distance, and using the map grid corresponding to the world coordinate label as the endpoint grid of the navigation task.

[0110] In one possible embodiment, the aforementioned drone-assisted unmanned vehicle positioning and navigation platform can receive the navigation task of transporting materials from point A to point B in an industrial park. Point A corresponds to the starting point coordinates of 113.123456°E, 23.654321°N, and point B corresponds to the ending point coordinates of 113.129876°E, 23.659876°N. The drone-assisted unmanned vehicle positioning and navigation platform can retrieve the 2.5D elevation map corresponding to the industrial park from the ground station's storage medium based on the target area indicated by the navigation task, and read the longitude, latitude, and elevation associated with each map grid.

[0111] Specifically, the aforementioned UAV-assisted unmanned vehicle positioning and navigation platform can sequentially compare the starting point coordinates with each of the aforementioned world coordinate labels. When a world coordinate label exists that matches the starting point coordinates, the map grid corresponding to that world coordinate label is determined as the starting point grid. When no world coordinate labels have completely identical coordinates, the map grid corresponding to the world coordinate label with the smallest coordinate distance from the starting point coordinates is determined as the starting point grid. Similarly, the aforementioned UAV-assisted unmanned vehicle positioning and navigation platform can compare the ending point coordinates with each of the aforementioned world coordinate labels to determine the ending point grid.

[0112] Through the above processing, the navigation task endpoints in the world coordinate system can be mapped to specific map grids in the 2.5D elevation map, providing endpoint basis for generating the target path between the starting grid and the ending grid.

[0113] Optionally, the step of converting the coordinates of path points in the target path to world coordinates based on coordinate transformation parameters to obtain the navigation path in the world coordinate system further includes: sequentially extracting multiple path points from the target path and the path point coordinates of each path point in the point cloud map coordinate system; obtaining the rotation matrix and translation vector of the point cloud map coordinate system relative to the world coordinate system from the coordinate transformation parameters; performing rotation transformation on the coordinates of each path point based on the rotation matrix, and performing translation transformation on the rotated path point coordinates based on the translation vector to obtain the world coordinates corresponding to each path point; and combining the world coordinates according to the order of the path points in the target path to obtain the navigation path.

[0114] In this embodiment of the invention, the above arrangement order refers to the order in which each of the above path points points from the starting grid to the ending grid in the above target path. Maintaining this arrangement order can preserve the continuous relationship between the path points.

[0115] In one possible embodiment, the smoothed target path comprises approximately 280 sequentially arranged path points. The aforementioned UAV-assisted unmanned vehicle positioning and navigation platform can sequentially read the coordinates of each path point in the aforementioned point cloud map coordinate system according to the arrangement order of the path points, and retrieve the aforementioned rotation matrix and translation vector associated with and saved in the current 2.5D elevation map.

[0116] Specifically, the aforementioned UAV-assisted unmanned vehicle positioning and navigation platform can perform rotation transformation on the coordinates of each path point based on the aforementioned rotation matrix, so that the coordinate direction of the path point is consistent with the coordinate direction of the aforementioned world coordinate system; then, based on the aforementioned translation vector, it performs position translation on the rotated and transformed path point coordinates to obtain the longitude, latitude, and elevation of the corresponding path point in the aforementioned world coordinate system. The aforementioned UAV-assisted unmanned vehicle positioning and navigation platform can combine the aforementioned world coordinates according to the original arrangement order of each path point in the aforementioned target path to obtain the aforementioned navigation path composed of multiple world coordinate path points in sequence, and load the aforementioned navigation path into the unmanned vehicle navigation module.

[0117] Therefore, the above navigation path and the on-board real-time dynamic differential positioning information obtained during the autonomous vehicle's operation can use the same coordinate reference, providing consistent path data for subsequent deviation calculation and vehicle driving control.

[0118] like Figure 2As shown, this embodiment of the invention also provides a drone-assisted unmanned vehicle positioning and navigation device 200, which includes: The acquisition module 201 is used to acquire three-dimensional point cloud data collected by the UAV in the target area and the real-time dynamic differential positioning information of the UAV corresponding to the three-dimensional point cloud data, and generate an original three-dimensional point cloud map with positioning tags. The determination module 202 is used to determine the coordinate transformation parameters between the point cloud map coordinate system and the world coordinate system based on the correspondence between the point cloud map coordinates in the original three-dimensional point cloud map and the world coordinates represented by the real-time dynamic differential positioning information of the UAV. The generation module 203 is used to perform point cloud compression processing on the original three-dimensional point cloud map to generate a 2.5-dimensional elevation map represented by the point cloud map coordinate system, and to determine the world coordinate label corresponding to each map grid in the 2.5-dimensional elevation map based on the coordinate transformation parameters. The response module 204 is used to respond to the navigation task, retrieve the 2.5D elevation map, determine the starting grid and the ending grid according to the starting coordinates and ending coordinates corresponding to the navigation task and the world coordinate labels corresponding to each map grid, and generate the target path in the point cloud map coordinate system based on the elevation information corresponding to each map grid. The conversion module 205 is used to convert the coordinates of the path points in the target path into world coordinates based on the coordinate transformation parameters, so as to obtain the navigation path in the world coordinate system; The navigation module 206 is used to acquire the on-board real-time dynamic differential positioning information of the unmanned vehicle during its driving process, determine the lateral deviation and heading deviation based on the on-board real-time dynamic differential positioning information and the navigation path, and control the unmanned vehicle to drive along the navigation path based on the lateral deviation and the heading deviation.

[0119] Optionally, the aforementioned acquisition module 201 is used for: The drone is controlled to fly over the target area along a preset route, and multiple point cloud frames are continuously collected by the airborne lidar and the point cloud timestamps corresponding to each point cloud frame are recorded. The airborne real-time dynamic differential positioning module continuously acquires multiple sets of real-time dynamic differential positioning information of the UAV and records the positioning timestamps corresponding to each set of real-time dynamic differential positioning information of the UAV. The real-time dynamic differential positioning information of the UAV includes longitude, latitude, elevation and positioning accuracy factor. Based on the temporal correspondence between each point cloud timestamp and each positioning timestamp, determine the real-time dynamic differential positioning information of the UAV corresponding to each point cloud frame; Each point cloud frame is associated with the corresponding UAV real-time dynamic differential positioning information, and the UAV real-time dynamic differential positioning information is used as the positioning tag of the corresponding point cloud frame to obtain multiple point cloud frames with positioning tags. The original 3D point cloud map is generated based on multiple point cloud frames with location tags.

[0120] Optionally, the aforementioned determining module 202 is used for: Select a first reference point and a second reference point with intervals set from the original 3D point cloud map; Obtain the point cloud map coordinates of the first reference point and the second reference point in the point cloud map coordinate system, respectively; Based on the corresponding real-time dynamic differential positioning information of the UAV, the world coordinates of the first reference point and the second reference point in the world coordinate system are determined. Based on the point cloud map coordinates of the first reference point and the second reference point, determine the first direction vector in the point cloud map coordinate system; Based on the world coordinates of the first reference point and the second reference point, determine the second direction vector in the world coordinate system; Based on the rotation relationship between the first direction vector and the second direction vector, determine the rotation matrix of the point cloud map coordinate system relative to the world coordinate system; Based on the rotation matrix, the point cloud map coordinates of any of the reference points are rotated and transformed, and the translation vector is determined according to the coordinate difference between the rotated point cloud map coordinates and the corresponding world coordinates. The rotation matrix and the translation vector are used as the coordinate transformation parameters.

[0121] Optionally, the above-mentioned generation module 203 is used for: The original 3D point cloud map is subjected to height threshold filtering to obtain ground and near-ground point clouds; According to the preset grid size, the ground and near-ground point cloud are projected onto the corresponding map grid; The 2.5D elevation map is generated based on the maximum, minimum, and average elevation values ​​of the point cloud points within each map grid.

[0122] Optionally, the above-mentioned generation module 203 is also used for: Obtain the grid coordinates of each map grid in the point cloud map coordinate system; Based on the coordinate transformation parameters, rotation and translation transformations are performed on the coordinates of each grid to obtain the global coordinates of each map grid in the world coordinate system. The longitude, latitude, and elevation corresponding to each of the global coordinates are used as the world coordinate labels for the corresponding map grid.

[0123] Optionally, the above response module 204 is used for: In response to the navigation task, the starting coordinates and ending coordinates corresponding to the navigation task are obtained, and the starting coordinates and ending coordinates are both located in the world coordinate system; Retrieve the 2.5D elevation map corresponding to the target area and read the world coordinate labels corresponding to each map grid. The starting point coordinates are matched with each of the world coordinate labels, and the map grid corresponding to the world coordinate label that matches the starting point coordinates is determined as the starting point grid. The endpoint coordinates are matched with each of the world coordinate labels, and the map grid corresponding to the world coordinate label that matches the endpoint coordinates is determined as the endpoint grid.

[0124] Optionally, the above-mentioned conversion module 205 is used for: Extract multiple path points and the path point coordinates of each path point in the point cloud map coordinate system sequentially from the target path; Obtain the rotation matrix and translation vector of the point cloud map coordinate system relative to the world coordinate system from the coordinate transformation parameters; Based on the rotation matrix, the coordinates of each path point are rotated and transformed, and based on the translation vector, the coordinates of the rotated path points are translated and transformed to obtain the world coordinates corresponding to each path point. The navigation path is obtained by combining the world coordinates according to the order in which the path points are arranged in the target path.

[0125] like Figure 3 As shown, this embodiment of the invention also provides an electronic device 300, including a processor, which can execute any of the above-mentioned drone-coordinated unmanned vehicle positioning and navigation methods.

[0126] Specifically, it includes a processor 301 and a memory 302, as well as a computer program stored in the memory 302 and capable of running on the processor 301, which executes a drone-cooperative unmanned vehicle positioning and navigation method, wherein: The processor 301 executes the calculator program for the UAV-coordinated unmanned vehicle positioning and navigation method stored in the memory 302, and performs the following steps: Acquire 3D point cloud data collected by the UAV in the target area and the real-time dynamic differential positioning information of the UAV corresponding to the 3D point cloud data, and generate an original 3D point cloud map with positioning tags. Based on the correspondence between the point cloud map coordinates in the original 3D point cloud map and the world coordinates represented by the real-time dynamic differential positioning information of the UAV, the coordinate transformation parameters between the point cloud map coordinate system and the world coordinate system are determined. The original 3D point cloud map is compressed to generate a 2.5D elevation map represented by the point cloud map coordinate system. The world coordinate labels corresponding to each map grid are determined in the 2.5D elevation map based on the coordinate transformation parameters. In response to the navigation task, the 2.5D elevation map is retrieved, and the starting grid and ending grid are determined according to the starting coordinates and ending coordinates corresponding to the navigation task and the world coordinate labels corresponding to each map grid. The target path in the point cloud map coordinate system is generated based on the elevation information corresponding to each map grid. Based on the coordinate transformation parameters, the coordinates of the path points in the target path are converted into world coordinates to obtain the navigation path in the world coordinate system; The system acquires real-time dynamic differential positioning information of the unmanned vehicle during its operation, determines lateral deviation and heading deviation based on the real-time dynamic differential positioning information and the navigation path, and controls the unmanned vehicle to travel along the navigation path based on the lateral deviation and the heading deviation.

[0127] Optionally, the processor 301 executes the process of acquiring the 3D point cloud data collected by the UAV in the target area and the real-time dynamic differential positioning information of the UAV corresponding to the 3D point cloud data, and generating an original 3D point cloud map with positioning tags, including: The drone is controlled to fly over the target area along a preset route, and multiple point cloud frames are continuously collected by the airborne lidar and the point cloud timestamps corresponding to each point cloud frame are recorded. The airborne real-time dynamic differential positioning module continuously acquires multiple sets of real-time dynamic differential positioning information of the UAV and records the positioning timestamps corresponding to each set of real-time dynamic differential positioning information of the UAV. The real-time dynamic differential positioning information of the UAV includes longitude, latitude, elevation and positioning accuracy factor. Based on the temporal correspondence between each point cloud timestamp and each positioning timestamp, determine the real-time dynamic differential positioning information of the UAV corresponding to each point cloud frame; Each point cloud frame is associated with the corresponding UAV real-time dynamic differential positioning information, and the UAV real-time dynamic differential positioning information is used as the positioning tag of the corresponding point cloud frame to obtain multiple point cloud frames with positioning tags. The original 3D point cloud map is generated based on multiple point cloud frames with location tags.

[0128] Optionally, the processor 301 executes the correspondence between the point cloud map coordinates in the original 3D point cloud map and the world coordinates represented by the real-time dynamic differential positioning information of the UAV, and determines the coordinate transformation parameters between the point cloud map coordinate system and the world coordinate system, including: Select a first reference point and a second reference point with intervals set from the original 3D point cloud map; Obtain the point cloud map coordinates of the first reference point and the second reference point in the point cloud map coordinate system, respectively; Based on the corresponding real-time dynamic differential positioning information of the UAV, the world coordinates of the first reference point and the second reference point in the world coordinate system are determined. Based on the point cloud map coordinates of the first reference point and the second reference point, determine the first direction vector in the point cloud map coordinate system; Based on the world coordinates of the first reference point and the second reference point, determine the second direction vector in the world coordinate system; Based on the rotation relationship between the first direction vector and the second direction vector, determine the rotation matrix of the point cloud map coordinate system relative to the world coordinate system; Based on the rotation matrix, the point cloud map coordinates of any of the reference points are rotated and transformed, and the translation vector is determined according to the coordinate difference between the rotated point cloud map coordinates and the corresponding world coordinates. The rotation matrix and the translation vector are used as the coordinate transformation parameters.

[0129] Optionally, the processor 301 performs point cloud compression processing on the original 3D point cloud map to generate a 2.5D elevation map represented by the point cloud map coordinate system, including: The original 3D point cloud map is subjected to height threshold filtering to obtain ground and near-ground point clouds; According to the preset grid size, the ground and near-ground point cloud are projected onto the corresponding map grid; The 2.5D elevation map is generated based on the maximum, minimum, and average elevation values ​​of the point cloud points within each map grid.

[0130] Optionally, the processor 301 performs the process of determining the world coordinate labels corresponding to each map grid in the 2.5D elevation map based on the coordinate transformation parameters, including: Obtain the grid coordinates of each map grid in the point cloud map coordinate system; Based on the coordinate transformation parameters, rotation and translation transformations are performed on the coordinates of each grid to obtain the global coordinates of each map grid in the world coordinate system. The longitude, latitude, and elevation corresponding to each of the global coordinates are used as the world coordinate labels for the corresponding map grid.

[0131] Optionally, the processor 301 executes the navigation task-responsive method, retrieves the 2.5D elevation map, and determines the starting and ending grids based on the starting and ending coordinates corresponding to the navigation task and the world coordinate labels corresponding to each map grid, including: In response to the navigation task, the starting coordinates and ending coordinates corresponding to the navigation task are obtained, and the starting coordinates and ending coordinates are both located in the world coordinate system; Retrieve the 2.5D elevation map corresponding to the target area and read the world coordinate labels corresponding to each map grid. The starting point coordinates are matched with each of the world coordinate labels, and the map grid corresponding to the world coordinate label that matches the starting point coordinates is determined as the starting point grid. The endpoint coordinates are matched with each of the world coordinate labels, and the map grid corresponding to the world coordinate label that matches the endpoint coordinates is determined as the endpoint grid.

[0132] Optionally, the processor 301 performs the step of converting the coordinates of path points in the target path into world coordinates based on the coordinate transformation parameters to obtain a navigation path in the world coordinate system, including: Extract multiple path points and the path point coordinates of each path point in the point cloud map coordinate system sequentially from the target path; Obtain the rotation matrix and translation vector of the point cloud map coordinate system relative to the world coordinate system from the coordinate transformation parameters; Based on the rotation matrix, the coordinates of each path point are rotated and transformed, and based on the translation vector, the coordinates of the rotated path points are translated and transformed to obtain the world coordinates corresponding to each path point. The navigation path is obtained by combining the world coordinates according to the order in which the path points are arranged in the target path.

[0133] This invention also provides a computer-readable storage medium storing a computer program. When executed by a processor, the computer program implements the various processes of the drone-assisted unmanned vehicle positioning and navigation method or the application-side drone-assisted unmanned vehicle positioning and navigation method provided in this invention, and achieves the same technical effect. To avoid repetition, it will not be described again here.

[0134] Those skilled in the art will understand that implementing all or part of the processes in the above embodiments can be done by a computer program instructing related hardware, and can be stored in a computer-readable storage medium. When executed, the program can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0135] The above description discloses only preferred embodiments of the present invention and should not be construed as limiting the scope of the present invention. Therefore, equivalent variations made in accordance with the claims of the present invention are still within the scope of the present invention.

Claims

1. A method for unmanned vehicle positioning and navigation in collaboration with unmanned aerial vehicles, characterized in that, include: Acquire 3D point cloud data collected by the UAV in the target area and the real-time dynamic differential positioning information of the UAV corresponding to the 3D point cloud data, and generate an original 3D point cloud map with positioning tags. Based on the correspondence between the point cloud map coordinates in the original 3D point cloud map and the world coordinates represented by the real-time dynamic differential positioning information of the UAV, the coordinate transformation parameters between the point cloud map coordinate system and the world coordinate system are determined. The original 3D point cloud map is compressed to generate a 2.5D elevation map represented by the point cloud map coordinate system. The world coordinate labels corresponding to each map grid are determined in the 2.5D elevation map based on the coordinate transformation parameters. In response to the navigation task, the 2.5D elevation map is retrieved, and the starting grid and ending grid are determined according to the starting coordinates and ending coordinates corresponding to the navigation task and the world coordinate labels corresponding to each map grid. The target path in the point cloud map coordinate system is generated based on the elevation information corresponding to each map grid. Based on the coordinate transformation parameters, the coordinates of the path points in the target path are converted into world coordinates to obtain the navigation path in the world coordinate system; The system acquires real-time dynamic differential positioning information of the unmanned vehicle during its operation, determines lateral deviation and heading deviation based on the real-time dynamic differential positioning information and the navigation path, and controls the unmanned vehicle to travel along the navigation path based on the lateral deviation and the heading deviation.

2. The unmanned vehicle positioning and navigation method in collaboration with drones as described in claim 1, characterized in that, The process of acquiring 3D point cloud data collected by the UAV in the target area and the corresponding real-time dynamic differential positioning information of the UAV, and generating an original 3D point cloud map with positioning tags, includes: The drone is controlled to fly over the target area along a preset route, and multiple point cloud frames are continuously collected by the airborne lidar and the point cloud timestamps corresponding to each point cloud frame are recorded. The airborne real-time dynamic differential positioning module continuously acquires multiple sets of real-time dynamic differential positioning information of the UAV and records the positioning timestamps corresponding to each set of real-time dynamic differential positioning information of the UAV. The real-time dynamic differential positioning information of the UAV includes longitude, latitude, elevation and positioning accuracy factor. Based on the temporal correspondence between each point cloud timestamp and each positioning timestamp, determine the real-time dynamic differential positioning information of the UAV corresponding to each point cloud frame; Each point cloud frame is associated with the corresponding UAV real-time dynamic differential positioning information, and the UAV real-time dynamic differential positioning information is used as the positioning tag of the corresponding point cloud frame to obtain multiple point cloud frames with positioning tags. The original 3D point cloud map is generated based on multiple point cloud frames with location tags.

3. The unmanned vehicle positioning and navigation method in collaboration with drones as described in claim 1, characterized in that, The determination of coordinate transformation parameters between the point cloud map coordinate system and the world coordinate system based on the correspondence between the point cloud map coordinates in the original 3D point cloud map and the world coordinates represented by the real-time dynamic differential positioning information of the UAV includes: Select a first reference point and a second reference point with intervals set from the original 3D point cloud map; Obtain the point cloud map coordinates of the first reference point and the second reference point in the point cloud map coordinate system, respectively; Based on the corresponding real-time dynamic differential positioning information of the UAV, the world coordinates of the first reference point and the second reference point in the world coordinate system are determined. Based on the point cloud map coordinates of the first reference point and the second reference point, determine the first direction vector in the point cloud map coordinate system; Based on the world coordinates of the first reference point and the second reference point, determine the second direction vector in the world coordinate system; Based on the rotation relationship between the first direction vector and the second direction vector, determine the rotation matrix of the point cloud map coordinate system relative to the world coordinate system; Based on the rotation matrix, the point cloud map coordinates of any of the reference points are rotated and transformed, and the translation vector is determined according to the coordinate difference between the rotated point cloud map coordinates and the corresponding world coordinates. The rotation matrix and the translation vector are used as the coordinate transformation parameters.

4. The unmanned vehicle positioning and navigation method in collaboration with drones as described in claim 1, characterized in that, The step of performing point cloud compression processing on the original 3D point cloud map to generate a 2.5D elevation map represented by the point cloud map coordinate system includes: The original 3D point cloud map is subjected to height threshold filtering to obtain ground and near-ground point clouds; According to the preset grid size, the ground and near-ground point cloud are projected onto the corresponding map grid; The 2.5D elevation map is generated based on the maximum, minimum, and average elevation values ​​of the point cloud points within each map grid.

5. The unmanned vehicle positioning and navigation method in collaboration with drones as described in claim 4, characterized in that, The process of determining the world coordinate labels corresponding to each map grid in the 2.5D elevation map based on the coordinate transformation parameters includes: Obtain the grid coordinates of each map grid in the point cloud map coordinate system; Based on the coordinate transformation parameters, rotation and translation transformations are performed on the coordinates of each grid to obtain the global coordinates of each map grid in the world coordinate system. The longitude, latitude, and elevation corresponding to each of the global coordinates are used as the world coordinate labels for the corresponding map grid.

6. The unmanned vehicle positioning and navigation method in collaboration with drones as described in claim 1, characterized in that, In response to the navigation task, the 2.5D elevation map is retrieved, and the starting and ending grids are determined based on the starting and ending coordinates corresponding to the navigation task and the world coordinate labels corresponding to each map grid, including: In response to the navigation task, the starting coordinates and ending coordinates corresponding to the navigation task are obtained, and the starting coordinates and ending coordinates are both located in the world coordinate system; Retrieve the 2.5D elevation map corresponding to the target area and read the world coordinate labels corresponding to each map grid. The starting point coordinates are matched with each of the world coordinate labels, and the map grid corresponding to the world coordinate label that matches the starting point coordinates is determined as the starting point grid. The endpoint coordinates are matched with each of the world coordinate labels, and the map grid corresponding to the world coordinate label that matches the endpoint coordinates is determined as the endpoint grid.

7. The unmanned vehicle positioning and navigation method in collaboration with drones as described in claim 1, characterized in that, The process of converting the coordinates of path points in the target path to world coordinates based on the coordinate transformation parameters to obtain the navigation path in the world coordinate system includes: Extract multiple path points and the path point coordinates of each path point in the point cloud map coordinate system sequentially from the target path; Obtain the rotation matrix and translation vector of the point cloud map coordinate system relative to the world coordinate system from the coordinate transformation parameters; Based on the rotation matrix, the coordinates of each path point are rotated and transformed, and based on the translation vector, the coordinates of the rotated path points are translated and transformed to obtain the world coordinates corresponding to each path point. The navigation path is obtained by combining the world coordinates according to the order in which the path points are arranged in the target path.

8. A drone-assisted unmanned vehicle positioning and navigation device, characterized in that, include: The acquisition module is used to acquire three-dimensional point cloud data collected by the UAV in the target area and the real-time dynamic differential positioning information of the UAV corresponding to the three-dimensional point cloud data, and generate an original three-dimensional point cloud map with positioning tags. The determination module is used to determine the coordinate transformation parameters between the point cloud map coordinate system and the world coordinate system based on the correspondence between the point cloud map coordinates in the original 3D point cloud map and the world coordinates represented by the real-time dynamic differential positioning information of the UAV. The generation module is used to perform point cloud compression processing on the original 3D point cloud map to generate a 2.5D elevation map represented by the point cloud map coordinate system, and to determine the world coordinate labels corresponding to each map grid in the 2.5D elevation map based on the coordinate transformation parameters. The response module is used to respond to the navigation task, retrieve the 2.5D elevation map, determine the starting grid and ending grid according to the starting coordinates and ending coordinates corresponding to the navigation task and the world coordinate labels corresponding to each map grid, and generate the target path in the point cloud map coordinate system based on the elevation information corresponding to each map grid. The transformation module is used to convert the coordinates of path points in the target path into world coordinates based on the coordinate transformation parameters, so as to obtain the navigation path in the world coordinate system; The navigation module is used to acquire the onboard real-time dynamic differential positioning information of the unmanned vehicle during its driving process, determine the lateral deviation and heading deviation based on the onboard real-time dynamic differential positioning information and the navigation path, and control the unmanned vehicle to drive along the navigation path based on the lateral deviation and the heading deviation.

9. An electronic device, characterized in that, include: The system includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the steps of the unmanned vehicle positioning and navigation method in accordance with any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the unmanned vehicle positioning and navigation method in accordance with any one of claims 1 to 7.

Images