A Method for Generating Passable Areas of a Mobile Robot in a Non-Road Scene

By generating the maximum inline rectangle and merging algorithm in non-road scenarios, detection error and accuracy problems in the prior art are solved, and efficient passable area generation and path planning are achieved.

CN115790596BActive Publication Date: 2025-07-25HEFEI GEN SONG AUTOMATION TECH
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
CN202211407932.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-11-10
Publication Date
2025-07-25
Estimated Expiration
2042-11-10

AI Technical Summary

Technical Problem

The prior art passable area detection methods in non-road scenarios lack generalization, easily generate errors, and the detection accuracy decreases.

Method used

By obtaining the key point position of the scene raster map and the mobile robot trajectory, a passable area is generated using the maximum inline rectangle search and merging algorithm, and path planning is optimized in combination with image processing technology.

Benefits of technology

It improves the efficiency and accuracy of passable areas in non-road scenarios, reduces the number of areas, and avoids incorrect path planning for trajectory-free areas.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115790596B_ABST
    Figure CN115790596B_ABST
Patent Text Reader

Abstract

The present invention discloses a method for generating a passable area of a mobile robot in a non-road scenario. The method uses the mobile robot to perform scene mapping to obtain a scene grid map and the key pose points of the mobile robot trajectory during the mapping process. The key points refer to the starting point of the mobile robot trajectory and the turning points on the trajectory. Two adjacent key points are called a pair of key points. For each pair of key points, a maximum inscribed rectangle search is performed in the grid map to obtain the maximum passable rectangle in the area where each pair of key points is located. The maximum passable rectangles with the same direction and intersection are merged to form a passable rectangle for the straight path. All the passable rectangles for the straight paths constitute the passable area of the mobile robot. The present invention gradually generates the passable area by using the mobile robot trajectory during the mapping process, and merges the scattered passable rectangles through some constraint conditions to reduce the number of areas and improve the generation efficiency and usability of the passable area.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of path planning, and in particular to a method for generating a passable area for a mobile robot in a non-road scenario. Background Art

[0002] With the continuous development of robot technology, various types of robots have been gradually widely used in all walks of life, bringing great convenience to production and life. Before an AGV performs autonomous operations, it is necessary to establish a grid map of the working scenario (i.e., map building) and simultaneously plan the operation path. In order to improve the deployment efficiency of the AGV, it is necessary to improve the automation degree of processes such as map building and path planning. Generating a rectangular passable area using the map building data can more conveniently complete the path planning task.

[0003] Currently, the methods for detecting passable areas based on lidar are generally divided into two categories, namely the grid method and the raw point cloud method. The invention patent application CN110244321A published by the State Intellectual Property Office on September 17, 2019, proposes a method for detecting a passable area of a road based on a three-dimensional lidar. This method removes the points above a certain height above the lidar to obtain the interest points of the laser point cloud, and uses the RANSAC algorithm for ground segmentation to distinguish the ground point cloud and the obstacle point cloud, rasterize the obstacle point cloud, extract the data points closest to the vehicle in each grid, and combine these data points to be the boundary points of the passable area.

[0004] The method for detecting a passable area of a road disclosed in the invention patent application CN110244321A has more data settings that rely on experience, resulting in the lack of generalization of the system. With a slight perturbation of the data, it is easy to cause errors or even mistakes in the detection results of the system. At the same time, the accuracy of the passable area detection method mainly for roads will also decrease in non-road scenarios. Summary of the Invention

[0005] In view of the technical problems existing in the existing passable area detection methods, the present invention proposes a method for generating a passable area for a mobile robot in a non-road scenario.

[0006] A method for generating a passable area for a mobile robot in a non-road scenario includes the following steps:

[0007] Step 1, using the mobile robot to perform scene mapping to obtain a scene grid map and the key pose points of the mobile robot trajectory during the mapping process. The key points refer to the starting point of the mobile robot trajectory and the turning points on the trajectory;

[0008] Step 2: Define two adjacent key points as a pair of key points. For each pair of key points, perform a maximum inscribed rectangle search in the grid map to obtain the maximum passable rectangle in the area where each pair of key points is located.

[0009] Step 3: Merge the maximum passable rectangles that have the same direction and intersect to form a passable rectangle for the straight path. All the passable rectangles for the straight paths constitute the passable area of the mobile robot.

[0010] Further, in Step 2, the maximum passable rectangle in the area where each pair of key points is located is formed in the following way:

[0011] Take the line segment between two adjacent key points, and with its center point as the axis, rotate the line segment successively by θ, 2θ, 3θ,..., nθ to form a set of candidate lines.

[0012] Taking each candidate line as a reference, expand p grids outward in its parallel direction and perpendicular direction respectively. Using the expanded grids as boundaries, form two parallel lines and two perpendicular lines. Detect whether the two formed parallel lines and two perpendicular lines pass through the occupied grids. If they do not pass through, expand p grids outward again. When the expanded parallel line or perpendicular line passes through the occupied grid, stop the outward expansion in this direction. Stop the outward expansion until it cannot be expanded in all four directions to obtain the maximum inscribed rectangle of this candidate line.

[0013] Select the one with the largest space from the maximum inscribed rectangles of all candidate lines as the maximum passable rectangle in the area where this pair of key points is located.

[0014] Further, in Step 3, the passable rectangles for the straight paths are merged in the following way:

[0015] Judge whether the included angle between the two key point connection lines is less than the threshold, and judge whether the two maximum passable rectangles corresponding to the two pairs of key points intersect. If both conditions are satisfied, the maximum passable rectangles are merged.

[0016] First, merge two pairs of key points into one pair, and then according to the merged pair of key points, obtain the maximum passable rectangle in the area where they are located, that is, the merging is completed.

[0017] Further, two pairs of key points are merged into one pair in the following way:

[0018] Suppose the two key point connection lines are R1R2 and N1N2 respectively. Taking the line where R1R2 is located as the reference line L, project point N1 and point N2 onto the reference line L. Suppose points P1 and P2 are obtained respectively. Select the two outermost points from points R1, R2, P1, and P2 as the two key points of the merged pair of key points.

[0019] The present invention gradually generates a passable area by using the trajectory of a mobile robot during the mapping process, and restricts the generation range of the passable area through the mapping trajectory, avoiding the generation of a passable area in a traceless area, which may lead to incorrect routes in subsequent path planning. For scattered passable rectangles, some constraints are used to merge them, reducing the number of areas and improving the generation efficiency and usability of the passable area. Description of the Drawings

[0020] Figure 1 It is a flowchart of a method for generating a passable area for a mobile robot in a non-road scenario;

[0021] Figure 2 It is a schematic diagram of a candidate line for a pair of key points;

[0022] Figure 3 It is an expansion diagram of a candidate line;

[0023] Figure 4 It is a schematic diagram of the largest passable rectangle to be merged;

[0024] Figure 5 It is a schematic diagram of the passable rectangle of the straight-line path after merging;

[0025] Figure 6 It is a schematic diagram of the merging of key point pairs. Detailed Description of the Invention

[0026] The present invention will be further described in detail below in conjunction with the drawings and specific embodiments. The embodiments of the present invention are given for purposes of illustration and description, and are not intended to be exhaustive or to limit the invention to the disclosed form. Many modifications and variations will be apparent to those of ordinary skill in the art. The embodiments are chosen and described in order to best explain the principles of the invention and its practical application, and to enable those of ordinary skill in the art to understand the invention and design various embodiments with various modifications suitable for specific purposes.

[0027] Embodiment 1

[0028] A method for generating a passable area for a mobile robot in a non-road scenario, as Figure 1 shown, includes the following steps:

[0029] 1. Use a mobile robot to perform scene mapping to obtain a scene grid map and the key point poses of the mobile robot trajectory during the mapping process. The key points refer to the starting point of the mobile robot trajectory and the turning points on the trajectory.

[0030] The mobile robot moves in the scene, records the movement trajectory, and generates a grid map. Each grid in the grid map is either in an occupied state or in a free state. During the mapping process, the starting point of the mobile robot's trajectory and the turning points on the trajectory are listed as key points for generating the passable area.

[0031] 2. Define two adjacent key points as a pair of key points. For each pair of key points, perform a maximum inscribed rectangle search in the grid map to obtain the maximum passable rectangle in the area where each pair of key points is located.

[0032] The maximum passable rectangle can be obtained through existing image processing methods. In this embodiment, the maximum passable rectangle in the area where each pair of key points is located is formed as follows:

[0033] Take the line L between two adjacent key points 1 and 2, and with its center point as the axis, rotate the line successively by θ, 2θ, 3θ,..., nθ to form a set of candidate lines. Refer to Figure 2 , where n is the rotation parameter and can be freely set according to actual application requirements;

[0034] Using each candidate line as a reference, expand p grids in the parallel and perpendicular directions to it respectively. Use the expanded grids as boundaries to form two parallel lines and two perpendicular lines. Refer to Figure 3 , Figure 3 In

[0035] expand with the line L as a reference. Here, p is the expansion parameter and can be freely set. The smaller the value of p is set, the more accurate the generated maximum passable rectangle is, but at the same time, the expansion here increases the operation time. Therefore, the value of p is selected in combination with the hardware conditions and actual application requirements;

[0036] Detect whether the two parallel lines and two perpendicular lines formed pass through the occupied grids. If not, expand p grids again. When the expanded parallel or perpendicular line passes through the occupied grid, stop the expansion in that direction until no expansion is possible in all four directions, and obtain the maximum inscribed rectangle of this candidate line;

[0037] For those cases where the generation of the maximum passable rectangle of the key point pair fails due to various special reasons (such as a moving object passing by during mapping and being judged as an obstacle), it can be manually corrected.

[0038] 3. Merge the maximum passable rectangles that are in the same direction and intersect to form a passable rectangle for the straight path. Refer to Figure 4 , Figure 5 , and all the passable rectangles for the straight paths constitute the passable area of the mobile robot.

[0039] The merging of the maximum passable rectangles can be achieved through existing image processing methods. In this embodiment, the straight-line path passable rectangles are merged in the following manner:

[0040] Judge whether the angle between the two connecting lines of the key points is less than the threshold, and judge whether the two maximum passable rectangles corresponding to the two pairs of key points intersect. If both conditions are satisfied, the maximum passable rectangles are merged. Among them, whether the two maximum passable rectangles intersect can be carried out by using OPENCV, which belongs to the prior art and will not be elaborated here;

[0041] First, merge the two pairs of key points into one pair, and then, according to the merged pair of key points, obtain the maximum passable rectangle in the area where they are located, that is, the merging is completed;

[0042] When merging the two pairs of key points into one pair, the two outermost points can be directly selected. However, this method is also likely to cause the direction to gradually deviate during the continuous merging of key point pairs;

[0043] Therefore, in this embodiment, the two pairs of key points are merged into one pair in the following manner: Assume that the two connecting lines of the key points are R1R2 and N1N2 respectively. Taking the line where R1R2 is located as the reference line L, project the points N1 and N2 onto the reference line L. Refer to Figure 6 , and assume that the points P1 and P2 are obtained respectively. Select the two outermost points from the points R1, R2, P1, and P2 as the two key points of the merged key point pair. When merging subsequent key point pairs, the same method is adopted, which can ensure that the direction will not deviate continuously due to the merging of key point pairs.

[0044] Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art and related fields without creative efforts shall fall within the protection scope of the present invention.

Claims

1. A method for generating a passable area of a mobile robot in a non-road scenario, characterized in that It includes the following steps: Step 1: Use a mobile robot to perform scene mapping to obtain a scene grid map and the key poses of the mobile robot trajectory during the mapping process. The key points refer to the starting point of the mobile robot trajectory and the turning points on the trajectory; Step 2: Call two adjacent key points a pair of key points. Conduct a maximum inscribed rectangle search in the grid map for each pair of key points to obtain the maximum passable rectangle in the area where each pair of key points is located, which is specifically formed by the following method: Take the line connecting two adjacent key points, and with its center point as the axis, rotate the line successively by θ, 2θ, 3θ,..., nθ to form a set of candidate lines; Taking each candidate line as a reference, expand p grids outward in its parallel and perpendicular directions respectively. Use the expanded grids as the boundaries to form two parallel lines and two perpendicular lines. Detect whether the two formed parallel lines and two perpendicular lines pass through occupied grids. If not, expand p grids again. When the expanded parallel line or perpendicular line passes through an occupied grid, stop the expansion in this direction until no expansion is possible in all four directions, and obtain the maximum inscribed rectangle of this candidate line; Select the one with the largest space from the maximum inscribed rectangles of all candidate lines as the maximum passable rectangle in the area where this pair of key points is located; Step 3: Merge the maximum passable rectangles with the same direction and intersection to form a passable rectangle for the straight path. All the passable rectangles for the straight path constitute the passable area of the mobile robot.

2. The method for generating the passable area of a mobile robot in a non-road scenario according to claim 1, wherein, In the said Step 3, the passable rectangles for the straight path are merged by the following method: Judge whether the included angle between the two key point connection lines is less than the threshold, and judge whether the two maximum passable rectangles corresponding to the two pairs of key points intersect. If both conditions are met, the maximum passable rectangles are merged; First, merge two pairs of key points into one pair, and then according to the merged pair of key points, obtain the maximum passable rectangle in the area where they are located, that is, the merging is completed.

3. The method for generating the passable area of the mobile robot in the off-road scenario according to claim 2, wherein, Two pairs of key points are merged into one pair by the following method: Suppose the two key point connection lines are R1R2 and N1N2 respectively. Take the line where R1R2 is located as the reference line L, project the points N1 and N2 onto the reference line L, and suppose the points P1 and P2 are obtained respectively. Select the two outermost points from the points R1, R2, P1, and P2 as the two key points of the merged pair of key points.

Citation Information

Patent Citations

  • Road passable area detection method based on three-dimensional laser radar

    CN110244321A

  • Robot passable area determination method and device thereof, storage medium and robot

    CN113475976A

  • Covering moving operation method of robot

    CN114281076A