JPH0768488A - Method for detecting invasion prohibiting region of movable robot - Google Patents

Abstract

PURPOSE:To detect the region of the necessary minimum according to the shape of a member as the invasion prohibiting region even when the shape of the member is complicate by detecting the invasion prohibiting region of a robot from the information on the coordinates of the contour of the pattern of the prohibiting region. CONSTITUTION:The member pattern which is formed by the plane figure of the part higher than the collision limit of the member to be present in the working range of a robot and a masking pattern where the plane figure of the robot is approximately touches internally are generated. (step 2, and step 3). In addition, the member pattern is covered by this masking pattern to generated the prohibiting region pattern, and the region where this pattern is made the plane figure is detected as the invasion prohibiting region of the robot. The invasion prohibiting region is detected as the region which is expanded from the member pattern by the half width or by the radius of the masking pattern, and the region of the necessary minimum according to the shape is detected as the invasion prohibiting region irrespective of the member shape.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for detecting a forbidden area of a mobile robot, which detects in advance a range of collision with a member when the mobile robot performing welding, transportation, etc. moves.

[0002]

2. Description of the Related Art Conventionally, in a mobile robot such as a gantry robot used for welding, moving, etc., when performing automatic operation under computer control, for example, the range in which the torch tip of a manipulator moving in a suspended state collides with a member. Must be detected in advance as an intrusion prohibition area to avoid collision. The detection of the intrusion prohibition area is conventionally performed by a computer as described below.

That is, when there is a plate-like member 1 standing upright and a wing-like member 2 attached to this member 1 as shown in FIG. And numerical data of the height of the torch tip are given to the computer. Then, the computer sets the height of the tip of the torch as the collision critical height H, and by geometrical calculation, obtains a plane figure obtained by projecting portions of the members 1 and 2 in FIG. In addition, one rectangular area 3 in the same figure, which takes into consideration the size of the robot, is detected as an intrusion prohibited area. In FIG. 7, 1'and 2'denotes the members 1,
2 indicates a portion having a height H or more, and b indicates a half width or radius of the plane figure of the robot.

[0004]

In the conventional detection method, since the intrusion prohibition area is detected as a rectangular area, an unnecessarily wide area is easily detected as an intrusion prohibition area particularly for a member having a complicated shape. There is a problem that the minimum necessary area corresponding to the shape of the member cannot be detected as the intrusion prohibition area, and as a result, inconvenience occurs such that the degree of freedom of movement of the robot is unduly limited. Further, when a plurality of members are scattered in the work area and a plurality of intrusion-prohibited areas are present in the work area of the robot, it is not invented to make all of these areas the minimum necessary area and to detect them without omission. .

According to the present invention, even if the shape of the member is complicated,
The purpose is to detect the minimum necessary area corresponding to the shape as an intrusion prohibited area. Further, it is also an object to detect all of the plurality of intrusion prohibition areas as necessary minimum areas without omission.

[0006]

In order to achieve the above-mentioned object, in the method for detecting a prohibited area of a mobile robot according to the present invention, the collision criticality of the members existing in the working range of the mobile robot performing welding, transportation, etc. A member pattern of a binary image in which the above height portion is projected on a horizontal plane is generated, a masking pattern of a binary image for a robot symbol in which a horizontal plane projection figure of a robot is almost inscribed is generated, and each pixel of the member pattern is masked. A prohibited area pattern of a binary image in which a member pattern is overlaid on the center of the pattern is generated, and a robot intrusion prohibited area is detected from information of the contour line coordinates of the pattern.

Further, according to a second aspect of the present invention, each of the plurality of member patterns is overcoated with a mask pattern to generate a plurality of prohibited area patterns, and the pattern detection is repeated from the scanning start edge of the screen on which each prohibited area pattern is drawn. Every time a prohibited area pattern closest to the scanning start edge is detected by this pattern detection, the invasion prohibited area of the mobile robot is detected from the information of the contour coordinates of the detection pattern, and the pattern is erased from the screen, and a plurality of robot intrusions are detected. Detect forbidden areas.

[0008]

In the case of the detection method of the present invention configured as described above, the member pattern formed by the planar figure of the height portion above the collision limit of the member existing in the working range of the robot and the planar figure of the robot are almost A contacting masking pattern is generated. Furthermore, this masking pattern fills the member pattern to generate a prohibited area pattern,
An area having this pattern as a planar shape is detected as a robot intrusion prohibited area.

At this time, the invasion prohibited area is detected as an area which is wider than the member pattern by half the width or radius of the masking pattern, and the minimum necessary area corresponding to the shape is invaded regardless of the shape of the member. It is detected as a prohibited area. Then, when a plurality of members are scattered in the work area of the robot and a plurality of prohibited area patterns are generated, detection of an intrusion prohibited area based on a pattern closest to the scanning start end of the screen on which these patterns are drawn, The deletion of the prohibited area pattern from the screen after the detection is repeated, and a plurality of intrusion prohibited areas are detected without omission.

[0010]

EXAMPLES Examples will be described with reference to FIGS. 1 to 6. In this embodiment, the present invention is applied to a portal robot used for welding, and the two axes of the horizontal plane are the X and Y axes and the vertical axis is the Z axis. In addition, the coordinate values of the contour line of the three-dimensional shape of each welding member existing in the work range of the robot are registered in advance in the work line segment coordinate file. Then, the computer executes the processes of steps S1 to S6 of FIG. 1 to detect the intrusion prohibited area of the robot.

That is, in the first step S1,
The computer reads the coordinate values registered in the work line segment coordinate file and at the same time, the Z-axis coordinate value (height) of the torch tip of the manipulator, which is the minimum ground height of the moving part of the robot.
Is given. And since all the parts higher than the torch tip of each welding member may collide with the robot, the computer sets the Z axis coordinate value of the torch tip as the collision critical height, and slices each welding member at this height, Find the height above the collision criticality of each welded member.

Next, in step S2, the line segments that form the contour of the height of the welding member above the collision threshold are obtained,
These line segments are projected on the horizontal plane to generate a binary image member pattern in which the heights of the collision critical parts or more of the respective welding members are projected on the horizontal plane. It should be noted that these patterns are generated, for example, with the welding member portion as logic 1 and the other portions as logic 0, and become as shown in patterns a, b, and c of FIG. Next, in step S3, the computer generates a masking pattern M of a binary image for a robot symbol from the coordinate value of the robot, for example, a rectangular pattern of logic 1 in FIG. 3 in which the horizontal plane projection figure of the robot is almost inscribed.

Further, this masking pattern M shown in FIG.
The pixel P at the center point shaded with is the member pattern a, b,
A logical sum operation is performed on each pixel of c, and each member pattern a, b, c is overcoated with a masking pattern M to generate a prohibited area pattern of a binary image. In addition,
The prohibited area patterns based on the member patterns a, b, and c of FIG. 2 are as shown in patterns α, β, and γ of FIG.

Then, the image data of the screen on which each prohibited area pattern is drawn is registered in the binary image file of the intrusion prohibited area. Next, in step S4, the computer reads the image data registered in the binary image file of the intrusion prohibited area, and for example, the respective prohibited area patterns α, β,
The screen on which γ is drawn is expanded in memory.

When each pixel on this screen is sequentially scanned from the upper left corner and displayed on a CRT or the like, each pixel on the screen is detected in order to detect an intrusion prohibited area based on each prohibited area pattern on the computer or screen without omission. Logic 1 and 0 are checked in the scanning order from the pixel in the upper left corner of the screen, and the detection of the prohibited area pattern closest to the upper left corner of the screen that first becomes logical 1 is repeated. Then, for example, in the case of the first pattern detection on the screen of FIG. 4, the pattern α of the figure is detected.

Further, in order to obtain information on the contour line coordinates of the detected prohibited area pattern, the computer extracts the pixels of the detected contour line of the prohibited area pattern α as shown in FIG. 5, and the actual pixels of these pixels are extracted. Calculate the coordinate value of the position. At this time, if the coordinate value of each pixel of the contour line is used as the information of the contour line coordinate as it is, the amount of information to handle becomes enormous. Therefore, the coordinates of each node of the contour line are encoded by the Freeman chain code. This code is compressed and, for example, the coordinates (xs, y) of the start position Ps of the encoding of FIG.
The code information of Table 1 to which s) is added is obtained as information of the contour line coordinates.

[0017]

[Table 1]

Based on this information and the height of the collision criticality, the computer detects the intrusion prohibited area of the robot having the prohibited area pattern α as a plane shape. Further, the prohibited area pattern α on the memory where the processing is completed is erased by filling it with a logic 0, and each registered data of the binary image file of the intrusion prohibited area is corrected to the image data of the screen shown in FIG. 6, for example.

Next, in step S6, the computer determines the presence / absence of an unprocessed prohibited area pattern based on the presence / absence of a logic 1 pixel in the memory. If the unprocessed prohibited area pattern remains, step S6 returns YES. After passing, the process returns to step S4, and the processing from step S4 is repeated. By repeating this process, for example, information about the contour line coordinates of the prohibited area pattern β in FIG. 6 is obtained, and then information about the contour line coordinates of the prohibited area pattern γ shown in FIG. 6 is obtained, and all intrusion prohibited areas are detected. When all the intrusion prohibited areas are detected, NO is passed through step S6, and the processing of the computer ends.

In this case, no matter what shape each member is scattered in the work area of the robot, the width of the robot in which each intrusion prohibited area is set by the masking pattern M by simple binary image processing. It can be obtained accurately and without omission as a wide minimum necessary area.
It should be noted that the actual welding is performed after the movement route of the robot is corrected or changed based on each obtained intrusion prohibited area.

In the above embodiment, the case where a plurality of intrusion prohibition areas exist in the movement area of the robot has been described. However, when there is only one intrusion prohibition area, the processing of steps S5 and S6 of FIG. It suffices to omit one, detect one prohibited area pattern, and obtain information about the contour line coordinates of the pattern.

Further, the present invention can be applied to detection of an invasion prohibited area of a mobile robot used for various transportations other than welding, and in this case, the height of the collision criticality is arbitrary including 0 based on the shape of the robot. Of course, it may be set to the value of. Further, the shape of the masking pattern is not limited to the embodiment, and the masking pattern may be circular, for example.

[0023]

Since the present invention is configured as described above, it has the following effects. A member pattern a, b, c formed by a plane figure of a height portion above the collision limit of a member existing in the work range of the robot and a masking pattern M in which the plane figure of the robot is almost inscribed are generated. The member patterns a, b, c are filled up and the prohibited area patterns α, β, γ
Is generated, and a region having these patterns α, β, γ as a planar shape is detected as a robot intrusion prohibited region. At this time, the intrusion prohibited region is half the width of the masking pattern M than the member patterns a, b, c. Since it is detected as an area expanded by the radius, the minimum necessary area corresponding to the shape is detected as the intrusion prohibited area by a simple binary graphic process regardless of the shape of the member.

Further, when a plurality of members are scattered in the work area of the robot and a plurality of prohibited area patterns α, β, γ are generated, the scanning start end of the screen on which these patterns α, β, γ are drawn. The detection of the intrusion prohibition area based on the pattern closest to and the deletion of the prohibition area pattern after the detection from the screen are repeated, and multiple intrusion prevention areas are detected without omission. The intrusion prohibition area is accurately detected without any omission as the minimum necessary area.

[Brief description of drawings]

FIG. 1 is an explanatory diagram of processing steps of an embodiment of a method for detecting a prohibited area for a mobile robot according to the present invention.

FIG. 2 is an explanatory diagram of an example of a member pattern in a work area.

FIG. 3 is an explanatory diagram of an example of a masking pattern.

FIG. 4 is an explanatory diagram of an example of a prohibited area pattern.

FIG. 5 is an explanatory diagram of contour coordinate extraction of a prohibited area pattern.

FIG. 6 is an explanatory diagram of a screen in which one of the prohibited area patterns is deleted.

FIG. 7 is an explanatory diagram of a conventional method.

[Explanation of symbols]

 a, b, c member pattern M masking pattern α, β, γ prohibited area pattern

Continuation of front page (51) Int.Cl. ⁶ Identification code Office reference number FI Technical display area G06T 1/00

Claims (2)

[Claims] 1. A member pattern of a binary image generated by projecting on a horizontal plane a height portion of a member existing in a work range of a mobile robot for welding, carrying, etc., which is above a collision criticality, and a horizontal plane projection figure of the robot is generated. A binary image masking pattern for a robot symbol that is substantially inscribed is generated, and a center of the masking pattern is superimposed on each pixel of the member pattern to generate a prohibited area pattern of the binary image in which the member pattern is overcoated. A method of detecting an intrusion prohibited area of a mobile robot, comprising detecting an intrusion prohibited area of the robot from information on contour line coordinates of the prohibited area pattern. 2. A plurality of member patterns are overcoated with a mask pattern to generate a plurality of prohibited area patterns, and pattern detection is repeated from the scanning start edge of the screen on which the respective prohibited area patterns are drawn. Each time the prohibited area pattern closest to the scanning start edge is detected, the invasion prohibited area of the mobile robot is detected from the information of the contour coordinates of the detected pattern, and the pattern is erased from the screen. The method for detecting a forbidden area of a mobile robot according to claim 1, wherein