Search

CN-121994252-A - Three-dimensional real-time navigation method and device of robot and electronic equipment

CN121994252ACN 121994252 ACN121994252 ACN 121994252ACN-121994252-A

Abstract

The application is applicable to the technical field of artificial intelligence and provides a three-dimensional real-time navigation method, a device and electronic equipment of a robot, wherein the method comprises the steps of determining a local updating range of an operation scene corresponding to the robot based on Octree Occupancy grid map (Octree-based Occupancy GRID MAPPING, octomap) information according to real-time operation data of the robot, and observing Occupancy probability and attenuation coefficient corresponding to each three-dimensional grid to be updated in the local updating range; updating the occupation probability of each three-dimensional grid to be updated according to the observation occupation probability and the attenuation coefficient corresponding to each three-dimensional grid to be updated; and determining a real-time navigation path of the robot in a local updating range according to the updated running scene Octomap information and the A search algorithm. Therefore, the reliability, the safety and the calculation efficiency of the three-dimensional space navigation of the robot are improved.

Inventors

  • XU JIAKAI

Assignees

  • 深圳市速腾聚创科技有限公司

Dates

Publication Date
20260508
Application Date
20260410

Claims (13)

  1. 1. A method for three-dimensional real-time navigation of a robot, comprising: acquiring current three-dimensional point cloud data acquired by a laser radar in the robot, a current depth image acquired by a depth camera and a current linear speed corresponding to the robot; determining a local updating range of an operation scene corresponding to the robot based on the octree occupancy grid map Octomap information according to the current linear speed and the current depth image, wherein the operation scene Octomap information comprises information of a plurality of three-dimensional grids; Determining the observation occupation probability corresponding to each three-dimensional grid to be updated and the attenuation coefficient corresponding to each three-dimensional grid to be updated according to the local updating range, the current three-dimensional point cloud data and the current depth image, wherein each three-dimensional grid to be updated is located in the local updating range; updating the occupation probability of each three-dimensional grid to be updated according to the observed occupation probability corresponding to each three-dimensional grid to be updated and the attenuation coefficient corresponding to each three-dimensional grid to be updated so as to generate updated operation scene Octomap information; And determining a real-time navigation path of the robot in the local updating range according to the updated operation scene Octomap information and an A search algorithm.
  2. 2. The method of claim 1, wherein determining the local update range of the operation scene Octomap information corresponding to the robot according to the current line speed and the current depth image includes: acquiring a minimum updating radius, a dynamic obstacle prediction time window and an anti-collision safety buffer radius corresponding to the robot; determining a basic radius corresponding to the local updating range according to the minimum updating radius corresponding to the robot, the dynamic obstacle prediction time window, the anti-collision safety buffer radius and the current linear speed; determining the type of the obstacle contained in the current depth image in the basic radius according to the basic radius and the current depth image; determining a semantic radius corresponding to the local updating range according to the type of the obstacle contained in the basic radius of the current depth image; and determining the local updating range according to the basic radius and the semantic radius.
  3. 3. The method of claim 1, wherein the determining the observed occupancy probability for each three-dimensional grid to be updated and the attenuation coefficient for each three-dimensional grid to be updated according to the local update range, the current three-dimensional point cloud data, and the current depth image comprises: Voxelized is carried out on the current three-dimensional point cloud data in the local updating range so as to determine each three-dimensional grid to be updated; determining initial occupation probability corresponding to each three-dimensional grid to be updated and the observed occupation probability; Performing target detection on the current depth image in the local updating range to determine the type and the confidence of the obstacle corresponding to each three-dimensional grid to be updated; And determining attenuation coefficients corresponding to the three-dimensional grids to be updated according to the types and the confidence degrees of the barriers corresponding to the three-dimensional grids to be updated, and correcting initial occupation probabilities corresponding to the three-dimensional grids to be updated.
  4. 4. The method of claim 3, wherein determining the attenuation coefficient corresponding to each three-dimensional grid to be updated according to the type of the obstacle and the confidence level corresponding to each three-dimensional grid to be updated, and correcting the initial occupation probability corresponding to each three-dimensional grid to be updated, comprises: correcting the initial occupancy probability corresponding to the first three-dimensional grid to be updated to a first occupancy probability value and determining the attenuation coefficient corresponding to the first three-dimensional grid to be updated to be a first attenuation value under the condition that the obstacle type corresponding to the first three-dimensional grid to be updated is a dynamic obstacle and the confidence coefficient is larger than a confidence coefficient threshold, wherein the first three-dimensional grid to be updated is any three-dimensional grid to be updated; and under the condition that the obstacle type corresponding to the first three-dimensional grid to be updated is a transparent obstacle, correcting the initial occupancy probability corresponding to the first three-dimensional grid to be updated to a second occupancy probability value, and determining the attenuation coefficient corresponding to the first three-dimensional grid to be updated to be a second attenuation value, wherein the second occupancy probability value is smaller than the first occupancy probability value, and the second attenuation value is smaller than the first attenuation value.
  5. 5. The method of any of claims 1-4, wherein the updated operation scene Octomap information includes a current occupancy probability of each of the three-dimensional grids after updating, and the determining, according to the updated operation scene Octomap information and an a-search algorithm, a real-time navigation path of the robot in the local update range includes: Determining each obstacle node and each non-obstacle node in the local updating range according to the current occupation probability and the idle probability threshold corresponding to each three-dimensional grid; Determining an initial node and a target node of the robot in the local updating range according to the historical navigation path corresponding to the robot, each obstacle node and each non-obstacle node; Searching each non-obstacle node according to the A-search algorithm and a preset cost function to determine a real-time navigation path between the starting node and the target node, wherein the preset cost function is determined according to an actual cumulative cost function, a heuristic function, a slope cost function and a rolling cost function.
  6. 6. The method of claim 5, wherein the starting node is a last unblocked node of the historical navigation path within the local update range.
  7. 7. The method of claim 6, wherein the actual cumulative cost of the starting node is the actual cumulative cost of the last unblocked node when searching the starting node by the a-search algorithm.
  8. 8. The method of claim 5, wherein when searching for an effective neighborhood node corresponding to a current node being searched in a closed list according to the a-search algorithm and a preset cost function, determining a slope cost of the effective neighborhood node by: Determining the height difference and the horizontal distance between the current node and the effective neighborhood node, wherein the effective neighborhood node is a non-obstacle node which is adjacent to the current node and is not in a closed list; And determining the slope cost of the effective neighborhood node according to the height difference, the horizontal distance and the slope cost function.
  9. 9. The method of claim 5, wherein when searching for an effective neighbor node corresponding to a current node being searched in a closed list according to the a-search algorithm and a preset cost function, determining a rolling cost of the effective neighbor node by: Determining a course angle variation of the robot from the current node to the effective neighborhood node, wherein the effective neighborhood node is a non-obstacle node which is adjacent to the current node and is not in a closed list; and determining the rolling cost of the effective neighborhood node according to the course angle variation and the rolling cost function.
  10. 10. A three-dimensional real-time navigation device of a robot, comprising: The first acquisition module is used for acquiring current three-dimensional point cloud data acquired by a laser radar in the robot, a current depth image acquired by a depth camera and a current linear speed corresponding to the robot; the first determining module is configured to determine a local update range of operation scene Octomap information corresponding to the robot according to the current linear velocity and the current depth image, where the operation scene Octomap information includes information of a plurality of three-dimensional grids; The second determining module is configured to determine, according to the local updating range, the current three-dimensional point cloud data and the current depth image, an observed occupation probability corresponding to each three-dimensional grid to be updated and an attenuation coefficient corresponding to each three-dimensional grid to be updated, where each three-dimensional grid to be updated is located in the local updating range; The first updating module is configured to update the occupancy probability of each three-dimensional grid to be updated according to the observed occupancy probability corresponding to each three-dimensional grid to be updated and the attenuation coefficient corresponding to each three-dimensional grid to be updated, so as to generate updated operation scene Octomap information; And the third determining module is used for determining a real-time navigation path of the robot in the local updating range according to the updated operation scene Octomap information and an A-search algorithm.
  11. 11. An electronic device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the method of any of claims 1-9 when executing the computer program.
  12. 12. A computer readable storage medium storing a computer program, characterized in that the computer program when executed by a processor implements the method according to any one of claims 1-9.
  13. 13. A computer program product comprising a computer program which, when run on an electronic device, causes the electronic device to perform the method of any of claims 1-9.

Description

Three-dimensional real-time navigation method and device of robot and electronic equipment Technical Field The application belongs to the technical field of artificial intelligence, and particularly relates to a three-dimensional real-time navigation method and device for a robot and electronic equipment. Background In the field of robot three-dimensional space navigation, rapid three-dimensional space point-to-point path planning is a key navigation technology of flying robots or off-road robots with complex terrains, real-time path planning needs to avoid space static obstacles, dynamic obstacles and the like, and features such as real-time performance, shortest path and the like need to be considered. In the related art, most robots perform path planning based on a two-dimensional map or a 2.5-dimensional (with height information) cost map, but because the two-dimensional map lacks height information, the 2.5-dimensional cost map needs to compress the 3-dimensional point cloud to a two-dimensional plane (maximum height projection), so that three-dimensional space information such as suspended obstacles (e.g. overpasses), multi-layer structures (shelves) and the like cannot be comprehensively expressed, and more than 30% of space information is lost. Therefore, the two-dimensional map or the 2.5-dimensional cost map lacks three-dimensional space information, so that the three-dimensional space navigation cannot be adapted, and the route planning is performed on the basis, so that the rationality and the safety of the route are low. Disclosure of Invention The embodiment of the application provides a three-dimensional real-time navigation method, a device, electronic equipment, a storage medium and a computer program product of a robot, which can solve the problems that three-dimensional space navigation cannot be adapted due to the fact that a two-dimensional map or a 2.5-dimensional cost map lacks three-dimensional space information, and route planning of the robot is performed on the basis of the three-dimensional space information, and the rationality and the safety of a route are lower. In a first aspect, the embodiment of the application provides a three-dimensional real-time navigation method of a robot, which comprises the steps of obtaining current three-dimensional point cloud data collected by a laser radar in the robot, current depth images collected by a depth camera and current linear speeds corresponding to the robot, determining a local update range of an operation scene corresponding to the robot based on Occupancy grid map (Octree-based Occupancy map GRID MAPPING, octomap) information of the Octree according to the current linear speeds and the current depth images, wherein the operation scene Octomap information comprises information of a plurality of three-dimensional grids, determining observation Occupancy probabilities corresponding to the three-dimensional grids to be updated and attenuation coefficients corresponding to the three-dimensional grids to be updated according to the local update range, the current three-dimensional point cloud data and the current depth images, wherein the three-dimensional grids to be updated are located in the local update range, updating the Occupancy probabilities of the three-dimensional grids to be updated according to the observation Occupancy probabilities corresponding to the three-dimensional grids to be updated and the attenuation coefficients corresponding to the three-dimensional grids to be updated, generating updated operation scene Octomap information, and determining a real-time navigation path of the robot according to the updated operation scene Octomap and the search algorithm A. In a possible implementation manner of the first aspect, the determining, according to the current linear velocity and the current depth image, a local update range of information of the operation scene Octomap corresponding to the robot includes: Acquiring a minimum updating radius, a dynamic obstacle prediction time window and an anti-collision safety buffer radius corresponding to the robot; determining a basic radius corresponding to a local updating range according to a minimum updating radius, a dynamic obstacle prediction time window, an anti-collision safety buffer radius and a current linear speed corresponding to the robot; determining the type of the obstacle contained in the current depth image in the basic radius according to the basic radius and the current depth image; Determining a semantic radius corresponding to the local updating range according to the type of the obstacle contained in the basic radius of the current depth image; And determining a local update range according to the basic radius and the semantic radius. In some embodiments, in another possible implementation manner of the first aspect, determining the observed occupation probability corresponding to each three-dimensional grid to be updated and the attenuation coefficient