Mobile machine for adjusting safety margin and moving method thereof

Abstract

A mobile machine for adjusting a safety margin and a method of moving the same are provided. The moving method includes generating a point cloud representing surrounding objects of the mobile machine by sensing the surrounding objects, determining channel candidates between the surrounding objects by using the point cloud, selecting a target channel from among the channel candidates based on a planned path, setting a safety margin based on a width of the target channel, and generating a moving target based on the target channel. And moving along the planned path while performing collision avoidance based on the safety margin.

Description

[0001] Cross-references to related applications

[0002] This application is based on and claims priority to Korean Patent Application No. 10-2024-0122422 filed on September 9, 2024, and Korean Patent Application No. 10-2024-0129811 filed on September 25, 2024, the disclosures of which are incorporated herein by reference in their entirety. Technical Field

[0003] The methods and apparatus consistent with the example embodiments relate to mobile machines, and more specifically, to mobile machines configured to adjust safety margins and methods of moving mobile machines. Background Technology

[0004] In some cases, path planning for mobile machines can be performed using point clouds. Point clouds can comprise points collected using LiDAR (Light Detection and Ranging), image sensors, depth sensors, ultrasonic sensors, or combinations thereof. Point clouds can represent objects or environments in three-dimensional (3D) space using points. Mobile machines can move to a target location while simultaneously avoiding collisions with surrounding obstacles through path planning. Summary of the Invention

[0005] One or more example embodiments may at least solve the above-described problems and / or disadvantages, as well as other disadvantages not described above. Furthermore, example embodiments do not need to overcome the above-described disadvantages, and example embodiments may not overcome any of the above-described problems.

[0006] According to one aspect of this disclosure, a method for moving a mobile machine is provided, the method comprising: generating point clouds corresponding to a plurality of objects in a region surrounding the mobile machine; determining channel candidates based on the point clouds corresponding to the plurality of objects; selecting a target channel from the channel candidates based on a planned path; setting a safety margin based on the width of the target channel; and moving along the planned path based on the safety margin.

[0007] According to another aspect of this disclosure, a control method for a mobile machine is provided, the control method comprising: receiving point clouds corresponding to a plurality of objects in a region surrounding the mobile machine; determining channel candidates based on the point clouds corresponding to the plurality of objects; selecting a target channel from the channel candidates based on a planned path; and determining a safety margin for avoiding collisions based on the difference between the width of the mobile machine and the width of the target channel.

[0008] According to another aspect of this disclosure, a mobile machine is provided, comprising: one or more sensors configured to sense a plurality of objects in an area surrounding the mobile machine; one or more processors configured to: determine channel candidates based on point clouds corresponding to the plurality of objects, select a target channel from the channel candidates based on a planned path, and set a safety margin based on the width of the target channel; and a drive system configured to move the mobile machine along the planned path based on the safety margin. Attached Figure Description

[0009] The above and / or other aspects will become more apparent from the description of certain exemplary embodiments in conjunction with the accompanying drawings, wherein:

[0010] Figure 1 This is a diagram schematically illustrating an exemplary configuration related to collision avoidance driving of a mobile machine according to an embodiment;

[0011] Figure 2 This is a diagram illustrating exemplary operation for driving a mobile machine according to an embodiment;

[0012] Figure 3 This is a flowchart illustrating a method of moving a mobile machine according to an embodiment;

[0013] Figure 4 This is a diagram illustrating an example of the process of generating point clouds and planning paths according to an embodiment;

[0014] Figure 5 This is a diagram illustrating an example of the process for determining channel candidates according to an embodiment;

[0015] Figure 6 This is a diagram illustrating an example of the process for determining a target channel from channel candidates according to an embodiment;

[0016] Figure 7 This is a diagram illustrating an example of the process of determining a target channel using channel lines and sub-path lines;

[0017] Figure 8 This is a diagram illustrating an example of the process for determining the safety margin width according to an embodiment;

[0018] Figure 9 This is a diagram illustrating an example of setting a wide safety margin according to an embodiment;

[0019] Figure 10 This is a diagram illustrating an example of setting a narrow safety margin according to an embodiment;

[0020] Figure 11 This is a flowchart illustrating an example of detailed operation of a mobile machine for adjusting safety margins according to an embodiment;

[0021] Figure 12A and Figure 12B This is a diagram illustrating examples of point-based channel search and cluster-based channel search according to embodiments;

[0022] Figure 13 This is a diagram illustrating an example of the process for updating the width of a target channel according to an embodiment;

[0023] Figure 14 This is a diagram illustrating an example of a channel search process using all points in a point cloud without a search area, according to an embodiment.

[0024] Figure 15 This is a diagram illustrating an example of a channel search process using a partial point of a point cloud by employing a search area, according to an embodiment.

[0025] Figure 16 This is a diagram illustrating an example of the process of estimating safety margin data using a neural network model according to an embodiment;

[0026] Figure 17 This is a block diagram illustrating an example configuration of a control device for a mobile machine according to an embodiment;

[0027] Figure 18 This is a flowchart illustrating a control method for a mobile machine according to an embodiment; and

[0028] Figure 19 This is a block diagram illustrating an example configuration of a mobile machine according to an embodiment. Detailed Implementation

[0029] The detailed structural or functional descriptions below are provided as examples only, and various changes and modifications can be made to the embodiments. These examples are not to be construed as limiting to this disclosure and should be understood to include all changes, equivalents, and substitutions within the spirit and technical scope of this disclosure.

[0030] This document may use terms such as first, second, etc., to describe various components. Each of these terms is not used to define the nature, order, or sequence of the corresponding component, but only to distinguish the corresponding component from other components. For example, the first component may be referred to as the second component, and similarly, the second component may be referred to as the first component.

[0031] It should be noted that if a component is described as "connected", "coupled", or "joined" to another component, then a third component may be "connected", "coupled", and "joined" between the first and second components, but the first component may be directly connected, coupled, or joined to the second component.

[0032] Unless the context clearly indicates otherwise, the singular forms “a,” “an,” and “the” are intended to include the plural forms as well. It will be further understood that the terms “including / contains” and / or “including / contains” as used herein specify the presence of the stated features, integers, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0033] As used in this article, “at least one of A and B”, “at least one of A, B or C”, etc., each of which can include any one of the items listed together in a corresponding phrase, or all possible combinations thereof.

[0034] Unless otherwise defined, all terms used herein (including technical and scientific terms) shall have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure pertains. Terms (such as those defined in common dictionaries) shall be interpreted as having the meaning consistent with their meaning in the context of the relevant field and shall not be interpreted in an idealized or overly formal sense unless expressly defined herein.

[0035] In the following description, embodiments will be described in detail with reference to the accompanying drawings. When describing embodiments with reference to the accompanying drawings, the same reference numerals denote the same elements, and repeated descriptions associated with them are omitted.

[0036] Figure 1 This is a diagram schematically illustrating an exemplary configuration related to collision avoidance driving of a mobile machine according to an embodiment. Reference Figure 1 The mobile machine 110 can move to the target location 102 along the planned path 103 based on autonomous driving. For example, the mobile machine 110 may include various machines or devices that perform autonomous driving. For example, the mobile machine 110 may include, but is not limited to, robots, robotic vacuum cleaners, vehicles, airplanes, drones, trains, or personal mobility devices.

[0037] The mobile machine 110 can move to the target location 102 while sensing surrounding objects 105 using sensor 111 and performing collision avoidance with surrounding objects 105. Sensor 111 may include one or more sensor components. For example, sensor 111 may include, but is not limited to, light detection and ranging (LiDAR), image sensors, depth sensors, ultrasonic sensors, or combinations thereof.

[0038] The mobile machine 110 can generate a point cloud 106 representing surrounding objects 105 using sensors 111. The point cloud 106 can include a number of three-dimensional (3D) points. 3D points can represent 3D coordinates. For example, 3D points can represent the 3D coordinates of the environment surrounding the mobile machine 110. The mobile machine 110 can determine a local pose 104 using the 3D points. The local pose 104 can represent the position and / or orientation of the mobile machine 110. The local pose 104 can be used to obtain or set a planned path 103 or to perform collision avoidance. Surrounding objects 105 can include various objects, such as stationary or moving objects. Stationary objects can include, but are not limited to, walls, pillars, furniture, etc. Moving objects can include, but are not limited to, people, animals, or vehicles that act as obstacles to the movement of the mobile machine 110.

[0039] Mobile machine 110 can set a target location 102 and plan a route 103 using map 101. Mobile machine 110 can receive map 101 from another device different from mobile machine 110, or it can generate map 101 itself. Target location 102 can be the destination of mobile machine 110. For example, target location 102 can be the final location that mobile machine 110 attempts to reach or an intermediate location on the way to the final location.

[0040] The mobile machine 110 can set up a safe zone 112 based on the surrounding environment, including surrounding objects 105. The safe zone 112 can be used as a virtual wall to avoid collisions. The mobile machine 110 can set up a planned path 103 to prevent collisions between the safe zone 112 and surrounding objects 105, and can control the movement of the mobile machine 110.

[0041] According to an embodiment, the mobile machine 110 can adaptively or dynamically adjust the safety margin. The safety margin can be a margin used to ensure the safety zone 112. In example cases using a fixed safety margin, optimal movement may be difficult. For example, in an example case where the safety margin is too wide, the mobile machine 110 may determine that it may not be able to pass through a space, even though the space is sufficient for the mobile machine 110 to pass through on the planned path 103 formed by the surrounding objects 105. In this case, the planned path 103 may not be optimal, or the mobile machine 110 may not move along the planned path 103. In an example case where the safety margin is too narrow, the mobile machine 110 may move too close to the surrounding objects 105, and thus may reduce driving safety. The mobile machine 110 can achieve optimal movement to the target location 102 by optimizing the safety margin according to the surrounding environment.

[0042] Figure 2 This is a diagram illustrating exemplary operation for driving a mobile machine according to an embodiment. (Reference) Figure 2The mobile machine can perform operations including, but not limited to, path planning 210 and motion control 220. For example, the mobile machine can perform path planning to obtain a planned path, and can perform motion control to move the mobile machine based on the planned path. The mobile machine can travel or move along the planned path via motion control 220. For example, the mobile machine can perform motion control 220 by using a drive system. For example, the movement speed, movement direction, or a combination thereof can be controlled based on motion control 220. For example, the drive system may include components for achieving movement, such as power generation components (e.g., a motor), power transmission components, steering components, or drives.

[0043] According to embodiments, the mobile machine can also perform operations including but not limited to motion coding 240, ambient environment sensing 250, position estimation 260, and safety margin control 270. For example, the mobile machine can perform motion coding 240 based on motion control 220. For example, motion data can be generated in motion coding operation 240. The motion data can represent the motion actually generated by motion control 220. For example, motion coding 240 can be performed by sensing the driving status of the driving system. The mobile machine can perform ambient environment sensing 250 by using sensors. Point clouds can be generated based on ambient environment sensing 250.

[0044] The mobile device can perform position estimation 260 based on map data 230, motion coding 240, and surrounding environment sensing 250. The current position of the mobile device can be estimated based on position estimation 260. The mobile device can receive map data 230 from another device different from itself, or it can generate map data 230 itself. The current position based on position estimation 260 can be used for path planning 210.

[0045] The mobile machine can perform safety margin control 270 based on path planning 210 and / or ambient sensing 250. For example, the mobile machine can use the planned path and / or point cloud for safety margin control 270. According to an embodiment, the mobile machine can identify channels on the planned path and can set a safety margin based on the width of the channel. The mobile machine can move along the planned path based on the safety margin.

[0046] Figure 3 This is a flowchart illustrating a method of moving a mobile machine according to an embodiment. (Reference) Figure 3The mobile machine can receive a point cloud representing objects in the area surrounding it in operation 310, determine channel candidates between surrounding objects in operation 320 using the point cloud, select a target channel from the channel candidates based on a planned path in operation 330, and determine a safety margin for collision avoidance based on the width of the target channel in operation 340. However, this disclosure is not limited thereto, and therefore, the method of moving the mobile machine according to the embodiments is not limited to this. Figure 3 The operations and / or the sequence of operations shown are illustrated. For example, a method of moving a mobile machine may include one or more other operations. In an embodiment, the mobile machine may include a control device specifically for safety margin control, and operations 310 to 340 may be performed by that control device. However, this disclosure is not limited thereto. Thus, according to another embodiment, the mobile machine may include an integrated control device for performing autonomous driving including safety margin control, and operations 310 to 340 may be performed by that integrated control device.

[0047] In operation 310, the method may include receiving point clouds corresponding to one or more objects near the mobile machine. For example, the method may include receiving point clouds corresponding to multiple objects in an area surrounding the mobile machine. For example, the point cloud may be data generated by sensing surrounding objects such as walls or furniture.

[0048] In operation 320, the method may include measuring the distance between points in a point cloud, identifying consecutive points with a distance exceeding a threshold, and determining the space between consecutive points as channel candidates. Points may be grouped based on clustering. Points spaced apart from each other by a distance less than a threshold may be classified into the same point group. The space between adjacent point groups may be determined as channel candidates.

[0049] For example, a mobile machine can sense its surroundings in a direction of rotation (e.g., clockwise or counterclockwise) and can sense consecutive points. Consecutive points can have different distances. In the example case where the distance between two consecutive points exceeds a threshold, the space between the two consecutive points can be identified as a channel candidate. For example, after sensing a first point, a second point can be sensed. The first and second points can be consecutive points. In the example case where the distance between the first and second points is less than a threshold, the first and second points can be classified into the same group. In the example case where the distance between the first and second points is greater than or equal to a threshold, the first and second points can be classified into different groups. In this case, the space between the first and second points can be identified as a channel candidate.

[0050] In operation 330, the method may include selecting a channel intersecting the planned path from the channel candidates as the target channel. The planned path can be selected from one of the channel candidates, and one of the channel candidates can be determined as the target channel.

[0051] According to an embodiment, operation 330 may include determining a passage line traversing a passage region through a candidate passage, determining sub-path lines between path points on a planned path, and selecting a target passage based on the geometric relationship between the passage line and the sub-path lines. The passage region may be the region between each selected point from an adjacent group of points in the passage. In the example case where a first group of points is adjacent to a second group of points, the distance between a first point in the first group and a second point in the second group is the shortest among points in the first and second groups, and the region between the first and second points may be the passage region. The line connecting the first point to the second point may be a passage line. Path points may be determined at certain distances on the planned path. Lines connecting adjacent path points among the path points may be determined as sub-path lines.

[0052] Channel candidates can each have channel lines. Channel candidates can include a first channel candidate having a first channel line among the channel lines. In the example case where the first channel line intersects with a first sub-path line among the sub-path lines and the intersection point is within the area formed by the first channel line and the first sub-path line, the first channel candidate can be selected as the target channel.

[0053] A safety margin can be determined based on the difference between the width of the mobile machine and the width of the target channel. For example, operation 340 may include setting the width of the safety margin to correspond to the difference between the width of the mobile machine and the width of the target channel. In the example case where the mobile machine is projected as a rectangle from the bird's-eye view, the width direction can be determined to be perpendicular to the front direction of the mobile machine, and the length of the width direction of the mobile machine can be the width of the mobile machine.

[0054] The moving machine can control its speed based on a safety margin. For example, as the safety margin widens, the moving machine can increase its speed, and as the safety margin narrows, it can decrease its speed. For instance, the speed can be set proportional to the safety margin. In an example case where the difference between the width of the moving machine and the width of the target channel is large (e.g., greater than a threshold), the probability of a collision is low when the moving machine crosses the target channel. In this case, the moving machine's speed can be set high so that it can pass through the target channel quickly. In an example case where the difference between the width of the moving machine and the width of the target channel is small (e.g., less than a threshold), the probability of a collision is high when the moving machine crosses the target channel. In this case, the moving machine's speed can be set low so that it can pass through the target channel safely.

[0055] Safety margins can be set based on a safety zone around the mobile machine and a cost map around surrounding objects. Safety margins can be shared by the safety zone and / or the cost map. For example, a safety margin can be shared by the safety zone and the cost map. Figure 1 The security margin can be shared in one location, or it can be shared at a greater ratio by either the security zone or the cost map. According to another embodiment, the entire security margin can be distributed across either the security zone or the cost map.

[0056] Figure 4 This is a diagram illustrating an example of the process of generating point clouds and planning paths according to an embodiment. (Reference) Figure 4 The mobile machine 410 can sense surrounding objects 401 and 402 using sensor 411. For example, surrounding object 401 can be a wall, and surrounding object 402 can be a stationary object. As a result of the sensing, a point cloud 403 representing the surrounding objects 401 and 402 can be generated. The mobile machine 410 can set a safety margin by using the point cloud 403 and / or planning a path 404.

[0057] Figure 5 This is a diagram illustrating an example of the process for determining channel candidates according to an embodiment. (Refer to...) Figure 5 The mobile machine can use point cloud 502 to determine channel candidates between surrounding objects. For example, channel candidates may include, but are not limited to, first channel candidate T1, second channel candidate T2, third channel candidate T3, fourth channel candidate T4, and fifth channel candidate T5. Surrounding objects can correspond to... Figure 4 The surrounding objects 401 and 402, and the point cloud 502 can correspond to Figure 4 The point cloud 403. The mobile machine can select a target channel from the channel candidates based on the planned path 501. For example, the mobile machine can also select a channel that intersects with the planned path 501 (e.g., the third channel candidate T3) as the target channel.

[0058] The mobile machine can determine the space between consecutive points as channel candidates. The mobile machine can measure the distance between points in point cloud 502 and can identify consecutive points with a distance exceeding a threshold. For example, the mobile machine can sense surrounding objects in a specific rotational direction (e.g., clockwise or counterclockwise) and can sense consecutive points through this sensing. Consecutive points can have different distances. In the example case where the distance between two consecutive points exceeds a threshold, the space between the two consecutive points can be determined as a channel candidate.

[0059] For example, after sensing a first point, a second point can be sensed. The first and second points can be consecutive. In the example case where the distance between the first and second points is less than a threshold, the first and second points can be classified into the same group. In the example case where the distance between the first and second points is greater than or equal to the threshold, the first and second points can be classified into different groups. In this case, the space between the first and second points can be determined as channel candidates.

[0060] Figure 6 This is a diagram illustrating an example of the process for determining a target channel from channel candidates according to an embodiment. (Refer to...) Figure 6 The mobile machine can determine channel lines based on channel candidates. Channel lines can include, but are not limited to, a first channel line L1, a second channel line L2, a third channel line L3, a fourth channel line L4, and a fifth channel line L5. For example, the mobile machine can also determine channel lines (e.g., first channel line L1 to fifth channel line L5) that traverse a region across a channel candidate (e.g., first channel candidate T1 to fifth channel candidate T5). The traversed region can be the area between each selected point (e.g., first point 602 and second point 603) from adjacent point groups (e.g., a first point group including a first point 602 to the left of first channel candidate T1 and a second point group including a second point 603 to the right of first channel candidate T1) of each channel (e.g., one of the first channel candidate T1 to the fifth channel candidate T5).

[0061] The mobile machine can determine waypoints at regular distances on the planned path 601. However, this disclosure is not limited to this, and therefore, the mobile machine can determine waypoints at irregular distances on the planned path 601. For example, waypoints may include, but are not limited to, a first waypoint PT1, a second waypoint PT2, and a third waypoint PT3. The mobile machine can determine sub-path lines between waypoints (e.g., a first sub-path line P1 and a second sub-path line P2). The mobile machine can determine lines connecting adjacent waypoints (e.g., the first waypoint PT1 and the second waypoint PT2, or the second waypoint PT2 and the third waypoint PT3) to each other as sub-path lines. The mobile machine can select a target channel (e.g., a third channel candidate T3) from channel candidates (e.g., the first channel candidate T1 to the fifth channel candidate T5) based on the geometric relationship between the channel lines and the sub-path lines.

[0062] Figure 7 This is a diagram illustrating an example of the process of determining a target channel using channel lines and subpath lines. (See reference...) Figure 7The third channel line L3 of the third channel candidate T3 can intersect with the second sub-path line P2 at point 720, and the intersection point 720 can be in the region 710 formed by the third channel line L3 and the second sub-path line P2. For example, the region 710 can be a rectangle, and the points 711 and 712 at both ends of the third channel line L3 and the points 713 and 714 at both ends of the second sub-path line P2 can be on each side of the rectangle.

[0063] Figure 8 This is a diagram illustrating an example of the process for determining the safety margin width according to an embodiment. (See reference) Figure 8 The mobile machine can determine the safety margin width 830 based on the target channel width 810 and the mobile machine width 820. For example, the mobile machine can set the safety margin width 830 to correspond to the difference between the mobile machine width 820 and the target channel width 810. In the example case where the mobile machine is projected as a rectangle from a bird's-eye view, the mobile machine width 830 can be determined to be perpendicular to the front direction of the mobile machine.

[0064] The mobile machine width 820 can be fixed. The safety margin width 830 can be adaptively adjusted to suit the surroundings of the mobile machine (e.g., the target channel width 810). In the example case where the target channel width 810 increases, the safety margin width 830 can also increase, and in the example case where the target channel width 810 decreases, the safety margin width 830 can also decrease. The safety margin width 830 can be determined based on a safety zone set around the mobile machine and a cost map set around surrounding objects. For example, the safety margin can be determined by the safety zone and the cost map. Figure 1 The security margin can be shared in one location, or it can be shared at a greater ratio by either the security zone or the cost map. According to another embodiment, the entire security margin can be distributed across either the security zone or the cost map.

[0065] Figure 9 This is a diagram illustrating an example of setting a wide safety margin according to an embodiment, and Figure 10 This is a diagram illustrating an example of a narrow safety margin configuration according to an embodiment. (Refer to...) Figure 9The mobile machine 910 can traverse the first target channel 903 to move along the planned path 940. The first target channel 903 can have a relatively wide width. In this case, the mobile machine 910 can have a relatively wide safety margin. The safety margin can be shared by the safety zone 901 and / or the cost map 902. In the example case using a narrow safety margin, the mobile machine 910 may traverse the first target channel 903 close to surrounding objects. In this case, it may be difficult to deal with unexpected situations, and the safety of movement may be reduced. The mobile machine 910 can traverse the first target channel 903 stably by using a wide safety margin set based on the wide width of the first target channel 903.

[0066] refer to Figure 10 Mobile machine 1010 can pass through Figure 9 After passing through the first target channel 903, the mobile machine 1010 proceeds through the second target channel 1003 to continue moving along the planned path 1040. The second target channel 1003 may have a relatively narrow width. In this case, the mobile machine 1010 can set a relatively narrow safety margin. The safety margin can be shared by the safety zone 1002 and / or cost maps 1002 and 1005. In the example case using a wide safety margin, the mobile machine 1010 may not move along the planned path 1040 and may change its path if it is determined that the mobile machine 1010 can avoid passing through the second target channel 1003. In this case, the movement efficiency may be reduced. The mobile machine 1010 can pass through the second target channel 1003 by using a narrow safety margin set based on the narrow width of the second target channel 1003.

[0067] In the example scenario where the variable object 1004 suddenly appears and creates another narrow channel, the mobile machine 1010 can pass through the narrow channel by narrowing the safety margin. Alternatively, in the example scenario where the movement of the variable object 1004, which has already formed a narrow channel, widens the channel width, the mobile machine 1010 can stably pass through the channel by widening the safety margin.

[0068] The mobile machine 1010 can control its movement speed based on a safety margin. For example, as the safety margin widens, the mobile machine 1010 can increase its movement speed, such as... Figure 9 As shown, and as the safety margin narrows, the moving machine 1010 can reduce its moving speed, such as... Figure 10 As shown. For example, the movement speed can be set to be proportional to a safety margin.

[0069] Figure 11 This is a diagram illustrating an example of detailed operation of a mobile machine for adjusting safety margins according to an embodiment. (Reference) Figure 11In operation 1101, the mobile machine can load a map and determine a target location. For example, the mobile machine can receive the map from another device different from itself, or it can generate the map itself. The target location can be the final location the mobile machine is attempting to reach or an intermediate location on the way to the final location.

[0070] In operation 1102, the mobile machine can perform path planning. The mobile machine can perform position estimation to estimate its current position and can perform path planning based on the current position and the target position. The planned path can be generated or updated based on the path planning. In operation 1103, the mobile machine can move along the planned path.

[0071] In operation 1104, the mobile machine can sense its surrounding environment. As a result of the sensing, a point cloud representing the surrounding objects can be generated.

[0072] In operation 1105, the mobile machine can check whether the distance between consecutive points in the point cloud is greater than a threshold. If the distance is less than the threshold, the mobile machine can repeat operation 1102. If the distance is greater than the threshold, the mobile machine can repeat operation 1106.

[0073] In operation 1106, the mobile machine can store points and channel candidates. The mobile machine can identify the space between points with a distance greater than a threshold as channel candidates, and can store the channel candidates and the points associated with the channel candidates.

[0074] In operation 1107, the mobile machine can determine candidate path lines. In operation 1108, the mobile machine can determine path points on the planned path and sub-path lines between path points. For example, a predetermined number of path points can be determined on the planned path. For example, path points can be arranged at equal distances on the planned path. Sub-path lines connecting adjacent path points can be determined.

[0075] In operation 1109, the mobile machine can check whether a channel candidate meets the target channel conditions. Target channel conditions may include the channel line intersecting with a sub-path line and / or the intersection being within the area formed by the channel line and the sub-path line. In an example where the channel candidate does not meet the target channel conditions, operation 1102 can be performed again. In an example where the channel candidate meets the target channel conditions, operation 1110 can be performed.

[0076] In operation 1110, the mobile machine can identify a channel candidate as the target channel. In operation 1111, the mobile machine can determine the target channel width. In operation 1112, the mobile machine can determine a safety margin. For example, the mobile machine can determine a safety margin that is narrower than the target channel width and wider than the mobile machine width.

[0077] Figure 12A and Figure 12B This is a diagram illustrating examples of point-based channel search and cluster-based channel search according to embodiments. For example, Figure 12A An example of a mobile machine performing a point-based channel search 1201 is shown, and Figure 12B An example of a mobile machine performing a cluster-based channel search 1202 is shown. Figure 12B As shown, the first point group 1210 and the second point group 1220 can be located at opposite ends of the target channel. The moving machine can determine point clusters to identify the first point group 1210 and the second point group 1220.

[0078] refer to Figure 12A In the example case of performing point-based channel search 1201, the mobile machine can search for channel candidates based on the distance between points. For example, the first point 1211 and the second point 1221 can be consecutive points. In the example case where the first distance D1 between the first point 1211 and the second point 1221 exceeds a threshold, the mobile machine can determine the space between the first point 1211 and the second point 1221 as a channel candidate.

[0079] In the case of noise, such as at point 1212, a channel candidate may not be found using point-based channel search 1201. In the case where the second distance D2 between point 1221 and point 1212 is less than a threshold, no channel candidate may be found between point 1221 and point 1212. In such cases, the mobile machine may not use the optimal path.

[0080] refer to Figure 12B In the example case of performing cluster-based channel search 1202, the mobile machine can determine channel candidates by using a third distance D3 between the first center of the first point group 1210 and the second center of the search point group 1220. In the example case where the third distance D3 exceeds a threshold, the mobile machine can determine the space between the first point group 1210 and the second point group 1220 as a channel candidate. In the example case of performing cluster-based channel search 1202, channel candidates can be searched robustly against noise (such as the third point 1212).

[0081] According to an embodiment, the mobile machine can selectively perform point-based channel search 1201 and cluster-based channel search 1202. For example, the mobile machine can determine channel candidates based on the distance between the nearest first point 1211 and second point 1221 in the first point group 1210 and the second point group 1220, the distance between the first center of the first point group 1210 and the second center of the second point group 1220, or a combination thereof. For example, the mobile machine can analyze the noise level of the point cloud and can selectively perform point-based channel search 1201 and cluster-based channel search 1202 based on the noise level. In example cases where the noise level is below a threshold, point-based channel search 1201 can be performed, and in example cases where the noise level is above a threshold, cluster-based channel search 1202 can be performed.

[0082] Figure 13 This is a diagram illustrating an example of the process for updating the width of a target channel according to an embodiment. (See reference) Figure 13 The mobile machine 1310 can determine the target channel width of the target channel 1301 based on the distribution of points obtained by sensing the target channel 1301. Although the actual target channel width is fixed, the distribution of points can continuously change according to the attitude of the mobile machine 1310, so the target channel width value can be sensed at different times. For example, the target channel width can be sensed at a first time t1 using a first value TW1, at a second time t2 using a second value TW2, at a third time t3 using a third value TW3, and at the k-th time t... K Through the k-th value TW K Sensing.

[0083] Mobile device 1310 can continuously update the target channel width based on changes in the sensed value of the target channel width, or it can update the target channel width under update conditions. In the example case of updating the target channel width every time its value changes, inefficient operation may occur. Operational efficiency can be improved by updating the target channel width under update conditions.

[0084] According to an embodiment, the target channel width can be determined based on the value (e.g., a first value TW1 to a k-th value TW). K The distribution of the sensed values ​​is used to set update conditions. In the example case where a new value is sensed outside the distribution of previously sensed values, the target channel width can be updated. In the example case where the target channel width is sensed as a first value in a first pose of the mobile machine, the mobile machine can determine the target channel width as the first value. In the example case where the target channel width is sensed as a second value in a second pose of the mobile machine based on a change in the mobile machine's pose (e.g., movement), the mobile machine can determine the target channel width as either the first or the second value based on the difference between the first and the second values.

[0085] For example, the update condition could also include a difference from a reference value exceeding a threshold. For example, the reference value could be an initial value for the target channel width when it is found. For example, a first value TW1 could be the initial value, and in an example case where the difference between the first value TW1 and the sensed second value TW2 and sensed third value TW3 is less than the threshold, the target channel width may not be updated. The k-th value TW1 sensed after the first value TW1 and the third value TW3... K If the difference between the values ​​exceeds the threshold, the target channel width can be updated to the k-th value TW. K .

[0086] Figure 14 This is a diagram illustrating an example of a channel search process using all points in a point cloud without a search area, according to an embodiment. Figure 15 This is a diagram illustrating an example of a channel search process using a partial point of a point cloud through a search area, according to an embodiment.

[0087] refer to Figure 14 The mobile machine 1410 can sense the surrounding environment in a rotational direction (e.g., clockwise or counterclockwise) and can sense consecutive points through this sensing. Consecutive points can have certain distances between them. In an example case where the distance between two consecutive points exceeds a threshold, the space between the two consecutive points can be identified as a channel candidate.

[0088] The first channel candidate T1 to the sixth channel candidate T6 can be determined by sensing in a certain rotational direction. In the example case where there are no continuous sensing points, channels may not be available. The mobile machine 1410 can determine the third channel candidate T3 as the target channel. The mobile machine 1410 can update the first safety margin width SW1 to the second safety margin width SW2 and can move along the planned path. The channel width of the seventh channel T7 can be narrower than the second safety margin width SW2. Omission of the seventh channel T7 may result in the mobile machine 1410 either not passing through the seventh channel T7 or colliding with the seventh channel T7.

[0089] refer to Figure 15 Mobile machine 1510 can perform channel search using search area 1521. Mobile machine 1510 can set a region around the planned path 1520 as search area 1521. For example, mobile machine 1510 can set search area 1521 based on a third channel candidate T3 on the predetermined path 1520. For example, mobile machine 1510 can set search area 1521 to include the third channel candidate T3 based on the channel width of the third channel candidate T3.

[0090] Mobile machine 1510 can determine channel candidates from points in search region 1521 of the point cloud. Points in search region 1521 can be some of all points in the point cloud. Mobile machine 1510 can determine the seventh channel candidate T7 as the target channel from these points. Mobile machine 1510 can determine the seventh channel candidate T7 as the target channel on the planned path, update the first safety margin width SW1 to a third safety margin width SW3 based on the channel width of the seventh channel T7, and move along the planned path. Detection of the seventh channel T7 can be achieved by using the third safety margin width SW3 to allow mobile machine 1410 to pass through the seventh channel T7.

[0091] Figure 16 This is a diagram illustrating an example of the process of estimating safety margin data using a neural network model according to an embodiment. (Reference) Figure 16 The mobile machine can input point cloud 1601 and / or path data 1602 into neural network model 1610, execute neural network model 1610, and obtain safety margin data 1611. Path data 1602 may include planned paths. Safety margin data 1611 may include safety margins.

[0092] Neural network model 1610 can be pre-trained using large-scale training data to output safety margin data 1611 based on inputs of point cloud 1601 and / or path data 1602. For example, neural network model 1610 may include a point network, convolutional neural network (CNN), recurrent neural network (RNN), large language model (LLM), or a combination thereof. In the example case where neural network model 1610 includes a CNN, point cloud 1601 and path data 1602 can be converted into image data and can be input into the CNN. In the example case where neural network model 1610 includes an LLM, point cloud 1601 and path data 1602 can be converted into token data and can be input into neural network model 1610. Neural network model 1610 can estimate safety margins with high accuracy in environments where there is a lot of noise or obstacles are complexly set up.

[0093] Figure 17 This is a block diagram illustrating an example configuration of a control device for a mobile machine according to an embodiment. (Reference) Figure 17The control device 1700 can receive point cloud 1701 and path data 1702, and can generate safety margin data 1703 based on the point cloud 1701 and path data 1702. The control device 1700 may include one or more processors 1710 and a memory 1720. The memory 1720 may store a control program 1721. The control device 1700 can generate the safety margin data 1703 by using the control program 1721. According to an embodiment, the mobile machine may include a control device 1700 specifically for performing safety margin control. The control device 1700 may be provided as a plug-in to the mobile machine for performing... Figure 2 Path planning 210 and position estimation 260. The control unit 1700 can be provided as a plug-in to the mobile machine for performing... Figure 2 The safety margin is controlled at 270.

[0094] Figure 18 This is a flowchart illustrating a control method for a mobile machine according to an embodiment. (Reference) Figure 18 In operation 1810, the mobile machine can generate a point cloud representing the surrounding objects by sensing the surrounding objects. In operation 1820, it can determine the channel candidates between the surrounding objects by using the point cloud. In operation 1830, it can select a target channel from the channel candidates based on the planned path. In operation 1840, it can set a safety margin based on the width of the target channel. In operation 1850, it can move along the planned path while performing collision avoidance based on the safety margin.

[0095] In operation 1820, the method may include measuring the distance between points in a point cloud, identifying consecutive points in the point cloud with a distance exceeding a threshold, and determining the space between consecutive points as channel candidates.

[0096] In operation 1830, the method may include selecting a channel from the channel candidates that intersects with the planned path as the target channel.

[0097] In operation 1830, the method may include determining a passage line through a passage candidate's traversal area, determining sub-path lines between waypoints on the planned path, and selecting a target passage based on the geometric relationship between the passage line and the sub-path lines.

[0098] The channel candidate may include a first channel candidate having a first channel line in the channel line, and selecting the target channel based on geometric relationships may include: selecting the first channel candidate as the target channel when the first channel line intersects with a first sub-path line in the sub-path line and the intersection point is within the area formed by the first channel line and the first sub-path line.

[0099] In operation 1840, the method may include setting the width of the safety margin to correspond to the difference between the width of the moving machine and the width of the target channel.

[0100] Mobile machines can control their movement speed based on safety margins.

[0101] Movement speed control can include increasing movement speed as the safety margin widens, and decreasing movement speed as the safety margin narrows.

[0102] The safety margin can be determined based on the safety zone set around the mobile machine and the cost map set around the surrounding objects.

[0103] According to an embodiment, the first point group and the second point group are located at both ends of the target channel, and in operation 1820, the method may include: determining channel candidates based on the distance between the closest first point and the second point in the first point group and the second point in the second point group, the distance between the first center of the first point group and the second center of the second point group, or a combination thereof.

[0104] The moving machine can determine the width of the target channel.

[0105] Determining the width of the target channel may include: determining the target channel width as a first value based on the target channel width being sensed as a first value in a first posture of the mobile machine, and determining the target channel width as a first value or a second value based on the difference between the first value and the second value based on the target channel width being sensed as a second value in a second posture of the mobile machine according to a change in the mobile machine's posture.

[0106] In operation 1820, the method may include setting a region around the planned path as a search region, and determining channel candidates from points in the search region in the point cloud.

[0107] Figure 19 This is a block diagram illustrating an example configuration of a mobile machine according to an embodiment. (Reference) Figure 19 The mobile machine 1900 may include one or more sensors 1910, one or more processors 1920, memory 1930, drive system 1940, storage device 1950, input / output (I / O) device 1960, and network interface 1970. These components may communicate with each other via communication bus 1980. However, this disclosure is not limited thereto, and therefore, according to another embodiment,

[0108] One or more sensors 1910 may include LiDAR, image sensors, depth sensors, ultrasonic sensors, or combinations thereof. One or more processors 1920 may execute instructions stored in memory 1930 or storage device 1950. When executed by one or more processors 1920, the instructions may cause the mobile machine 1900 to perform reference... Figures 1 to 18 The described operation. Memory 1930 may include a non-transitory computer-readable storage medium or a non-transitory computer-readable storage device. For example, one or more processors 1920 may be an integrated control device for performing autonomous driving, including safety margin control.

[0109] Memory 1930 may store instructions to be executed by one or more processors 1920, and may store related information when the mobile machine 1900 executes software and / or applications. Memory 1930 may store control program 1931. When at least a portion of control program 1931 is stored in memory 1930, reference... Figures 1 to 18 The described operations can be performed by the mobile machine 1900. The drive system 1940 may include components for achieving travel, such as power generation components (e.g., a motor), power transmission components, steering components, or drives.

[0110] Storage device 1950 may include a computer-readable storage medium or a computer-readable storage device. Storage device 1950 can store more information than memory 1930 for a longer period of time. For example, storage device 1950 may include a magnetic hard disk, optical disk, flash memory, floppy disk, or other non-volatile memory known in the art.

[0111] I / O device 1960 can receive input from a user using a keyboard and mouse in traditional input methods as well as in new input methods such as touch input, voice input, and image input. For example, I / O device 1960 may include a keyboard, mouse, touchscreen, microphone, or any other device that detects input from the user and transmits the detected input to mobile machine 1900. I / O device 1960 can provide output from mobile machine 1900 to the user through visual, auditory, or tactile channels. I / O device 1960 may include, for example, a display, touchscreen, speaker, vibration generator, or any other device that provides output to the user. Network interface 1970 can communicate with external devices via wired or wireless networks.

[0112] The units described herein can be implemented using hardware components, software components, and / or combinations thereof. The processing device can be implemented using one or more general-purpose or special-purpose computers, such as processors, controllers, and arithmetic logic units (ALUs), digital signal processors (DSPs), microcomputers, field-programmable gate arrays (FPGAs), programmable logic units (PLUs), microprocessors, or any other means capable of responding to and executing instructions in a defined manner. The processing device can run an operating system (OS) and one or more software applications running on the OS. The processing unit can also access, store, manipulate, process, and generate data in response to the execution of software. For simplicity, the description of the processing unit is used as the singular; however, those skilled in the art will understand that a processing unit can include multiple processing elements and various types of processing elements. For example, a processing unit can include multiple processors, or a single processor and a single controller. Furthermore, different processing configurations are possible, such as parallel processors.

[0113] Software may include computer programs, code, instructions, or some combination thereof, to independently or jointly instruct or configure a processing unit to operate as needed. Software and data may be stored in any type of machine, component, physical or virtual device, or computer storage medium or apparatus capable of providing instructions or data to or being interpreted by the processing unit. Software may also be distributed across network-coupled computer systems, enabling it to be stored and executed in a distributed manner. Software and data may be stored on one or more non-transitory computer-readable recording media.

[0114] The methods described in the examples above can be recorded in a non-transitory computer-readable medium containing program instructions to implement the various operations described above. The medium may also include data files, data structures, etc., alone or in combination with the program instructions. The program instructions recorded on the medium may be program instructions specifically designed and constructed for the purposes of the examples, or they may be program instructions known and available to those skilled in the art of computer software. Examples of non-transitory computer-readable media include magnetic media, such as hard disks, floppy disks, and magnetic tapes; optical media, such as CD-ROMs and DVDs; magneto-optical media, such as optical discs; and hardware devices specifically configured to store and execute program instructions, such as read-only memory (ROM), random access memory (RAM), flash memory, etc. Examples of program instructions include both machine code generated by a compiler and files containing higher-level code that can be executed by a computer using an interpreter.

[0115] The aforementioned device can act as one or more software modules to perform the operations described in the examples above, and vice versa.

[0116] As described above, although examples have been described with reference to limited accompanying drawings, those skilled in the art can apply various technical modifications and variations based on them. For example, according to some embodiments, suitable results can be achieved by performing the described techniques in a different order and / or by omitting, replacing, and / or supplementing components in the aforementioned systems, architectures, devices, or circuits in different ways. For example, in some embodiments, components in the aforementioned systems, architectures, devices, or circuits can be combined, replaced, and / or supplemented with other components or their equivalents.

[0117] Therefore, other embodiments are within the scope of the appended claims.

Claims

1. A method of moving a mobile machine, the method comprising: generating a point cloud corresponding to a plurality of objects in an area surrounding the mobile machine; determining a passage candidate based on the point cloud corresponding to the plurality of objects; selecting a target passage from the passage candidate based on a planned path; setting a safety margin based on a width of the target passage; and moving along the planned path based on the safety margin. determining the passage candidate comprises:

2. The method of claim 1, wherein, measuring distances between a plurality of points of the point cloud corresponding to the plurality of objects; identifying consecutive points of the plurality of points having distances exceeding a threshold; and determining a space between the consecutive points as the passage candidate. selecting the target passage comprises selecting a passage intersecting the planned path from the passage candidate as the target passage.

3. The method of claim 1, wherein, selecting the target passage comprises:

4. The method of claim 1, wherein, determining a passage line passing through a through area of the passage candidate; determining a sub-path line between path points on the planned path; and selecting the target passage based on a spatial relationship between the passage line and the sub-path line. the passage candidate comprises a first passage candidate having a first passage line of the passage line, 5. The method of claim 4, wherein, wherein selecting the target passage based on the spatial relationship comprises selecting the first passage candidate as the target passage based on the first passage line forming an intersection with a first sub-path line of the sub-path line and the intersection being within an area formed by the first passage line and the first sub-path line. setting the safety margin comprises setting a width of the safety margin based on a difference between a width of the mobile machine and a width of the target passage.

6. The method of claim 1, wherein, controlling a moving speed based on the safety margin.

7. The method of claim 1, further comprising: controlling the moving speed comprises increasing the moving speed as the safety margin widens and decreasing the moving speed as the safety margin narrows.

8. The method of claim 7, wherein, the safety margin is determined based on a safety area set around the mobile machine and a cost map set around the plurality of objects.

9. The method of claim 1, wherein, a first set of points is located at a first end of the target passage and a second set of points is located at a second end of the target passage, 10. The method of claim 1, wherein, wherein determining the passage candidate comprises determining the passage candidate based on a first distance between a first point of the first set of points and a second point of the second set of points that are closest to each other, a second distance between a first center of the first set of points and a second center of the second set of points, or a combination of the first distance and the second distance.

11. The method of claim 1, further comprising determining a width of the target passage. determining the width of the target passage comprises:

12. The method of claim 11, wherein, determining the width of the target passage as a first value based on the width of the target passage being sensed at a first pose of the mobile machine as the first value; and determining the width of the target passage as the first value or a second value based on a difference between the first value and the second value based on the width of the target passage being sensed at a second pose of the mobile machine as the second value according to a change in pose of the mobile machine. determining the passage candidate comprises:

13. The method of claim 1, wherein, ​ setting a first region around the planned path as a search region; and determining the passage candidate from points in the search region in the point cloud.

14. A control method of a mobile machine, the control method comprising: receiving a point cloud corresponding to a plurality of objects in a region around the mobile machine; determining a passage candidate based on the point cloud corresponding to the plurality of objects; selecting a target passage from the passage candidate based on a planned path; and determining a safety margin for collision avoidance based on a difference between a width of the mobile machine and a width of the target passage.

15. The control method according to claim 14, wherein determining the passage candidate comprises: measuring distances between a plurality of points of the point cloud corresponding to the plurality of objects; identifying consecutive points in the plurality of points having distances exceeding a threshold; and determining a space between the consecutive points as the passage candidate.

16. The control method according to claim 14, wherein selecting the target passage comprises selecting a passage intersecting the planned path from the passage candidate as the target passage.

17. The control method according to claim 14, wherein selecting the target passage comprises: determining a passage line passing through a through region of the passage candidate; determining a sub-path line between path points on the planned path; and selecting the target passage based on a spatial relationship between the passage line and the sub-path line.

18. The control method according to claim 14, wherein setting the safety margin comprises setting a width of the safety margin based on a difference between a width of the mobile machine and a width of the target passage.

19. The control method according to claim 14, wherein the safety margin is set based on a safety region set around the mobile machine and a cost map set around the plurality of objects.

20. A mobile machine comprising: one or more sensors configured to sense a plurality of objects in a region around the mobile machine; one or more processors configured to: determine a passage candidate based on a point cloud corresponding to the plurality of objects, select a target passage from the passage candidate based on a planned path; set a safety margin based on a width of the target passage; and a drive system configured to move the mobile machine along the planned path based on the safety margin.

Images