JP2890874B2 - Robot arm relative posture correction method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method of correcting a relative posture of a robot arm, and more particularly to a method of correcting a relative posture using a mark for positioning a relative posture of a robot when performing an assembly docking operation or the like.

[0002]

2. Description of the Related Art Conventionally, a three-dimensional mark has been considered as a mark of this kind, and a translational position shift and a rotational shift are detected based on a position and a way of seeing the three-dimensional mark. FIG. 4 and FIG.
It is sectional drawing and a front view showing a mark used for the conventional robot arm relative posture correction method shown in 264804. In the figure, a positioning mark 1 is provided with a step from a center circle 2 and is provided with an intermediate ring 3 formed by hollowing out a truncated cone and an outermost ring 4 provided on the front. As shown in FIG. 5, the center circle 2 is painted white, and a cross line 5 intersecting at the center is provided on the surface thereof. The cross line 5 is painted black so that its shape and center position are clear. Intermediate ring 3
Is painted white as in the case of the center circle 2. The intermediate ring 3 is provided with a black cut 6 located on the extension of the cross line 5. The outermost ring 4 is concentric with the center circle 2 and the intermediate ring 3, and its surface is painted white. The tapered walls 7, 8 formed by hollowing out in a truncated cone shape are painted black in order to make the outline of the center circle 2, the intermediate ring 3, and the outermost ring 4 clear.

The conventional positioning mark is configured as described above. When the optical axis of the camera coincides with the center axis of the mark, an image is taken at the center of the field of view of the camera as shown in FIG.
If they do not match, for example, the center position 17 of the cross line of the positioning mark and the intermediate ring and the outermost ring (center position 16) are imaged eccentrically as shown in FIG. 6, so that the image is taken as shown in FIG. In the case where the optical axis of the camera and the center axis of the mark are coincident with each other, but there is a rotational displacement, the gap of the intermediate ring is displaced from the predetermined position, It is said that the rotation should be corrected accordingly.

[0004]

The conventional method for correcting the relative posture of the robot arm has been described above, and the above-described conventional positioning mark and positioning method have the following disadvantages. (1) The mark itself has a complicated shape and color scheme, and a complicated program is required for automatic recognition by a computer image processing technique. (2) Further, even if the camera optical axis and the center axis of the positioning mark are aligned by the above-described position correction method, it does not guarantee that the position of the end effector actually acting on the object is aligned, For that purpose, it is necessary to strictly calibrate the relative relationship between the camera optical axis and the center axis of the end effector. This will be described with reference to the drawings. In FIG. 7, 10 is a camera, 20, 20 ′ are end effectors,
Reference numerals 30 and 30 ′ denote robot arms integrated with the end effectors 20 and 20 ′, respectively. The optical axis of the camera 10 is aligned with the center axis of the positioning mark 1. At this time, if the relative positional relationship between the optical axis of the camera and the center axis of the end effector is strictly calibrated, the end effector matches the position of the end effector working position 40 as shown at 20 ', but otherwise, for example, the end effector The position of the effector 20 deviates from the end effector operation position 40. This relative position can be changed by changing the camera lens,
In some cases, even if the focal point is changed, the center axis of the end effector is difficult to observe when attached to the robot, so that exact calibration is very difficult in practice. (3) As described above, after the camera optical axis is aligned with the center axis of the positioning mark, the rotational deviation of the axis is corrected, but the rotation angle is calculated from the difference between the center position of the cross line and the position of the cut in the intermediate ring. For this reason, it is difficult to increase the accuracy, and even a slight error in the rotation angle increases the error in the position of the end effector if the camera mounting arm is long. This will be described with reference to the drawings. FIG. 8 is a view seen from the direction of the optical axis of the camera, and 35 is a camera mounting arm. In the figure, the camera optical axis and the center axis of the positioning mark are aligned, but since the rotation angle is recognized as being shifted by θ from the measurement error of the position of the break of the intermediate ring, the end effector is originally located at the position of 20 ′. This shows that the place to come is at position 20. This shift amount increases as the distance between the center of the camera optical axis and the center axis of the end effector increases.

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and accurately matches the relative posture between a work target and an end effector by using a simple mark which can be easily automatically recognized by image processing technology. It is an object of the present invention to provide a method of correcting a relative posture of a robot arm that can perform the method.

[0006]

According to a method of correcting a relative posture of a robot arm according to the present invention, three or more marks are arranged on the same circumference centering on a position where an end effector is to act, and the end effector is provided. a television camera mounted on the robot arm shaft are integrated with, the robotic arm shaft is rotated by a predetermined amount by sequentially imaging the marks, the marks captured, the position in the same field of view of the television camera And calculate the relative attitude correction amount of the end effector based on the measurement results.
Then, the attitude of the end effector is determined by the relative attitude correction amount.
Is to be corrected .

[0007]

According to the robot arm relative posture correction method of the present invention, a plurality of marks arranged on the same circumference centering on the position where the end effector is to act are provided on the axis of the tip of the robot arm integrated with the end effector. in a television camera mounted sequentially captured by rotating the robot arm axis by a predetermined amount, the position where the mark is caught on the field of view of said television camera, a robot arm to calculate the relative position correction amount of the end-et Fekuta Since the relative posture is corrected, accurate relative posture correction can be performed with a simple mark, and a deviation between the center axis of the work target and the end effector can be detected, and the relative posture can be accurately matched.

[0008]

[Embodiment 1] FIG. 1 shows an embodiment of the present invention.
-4 is four positioning marks, 10 is camera, 2
0 is an end effector, 30 is a robot arm integrated with the end effector, 35 is a camera mounting arm for mounting the camera 10 to the robot arm, 4
Reference numeral 0 denotes an operation position where the end effector 20 operates.
It is installed at intervals of degrees. Each positioning mark has the shape of a truncated cone, and its side surface is colored black so that it can be clearly distinguished from the top surface and the background.

Next, the operation will be described with reference to the flowchart of FIG. First, in step S1, the robot arm is controlled to roughly position the end effector near its operating position. In step S2, the robot arm is rotated with the camera by 90 degrees around the axis of the end effector, and the marks 1-1 to 1-4 are sequentially captured by the camera. When the imaging is completed, the marks are rotated by -270 degrees to rotate the end effector. Return to the original position. In step S3, the amount of deviation of the end effector from the target position is calculated from the image data of the mark. In step S4, the displacement is evaluated. If it is within the allowable range, the positioning is completed. Otherwise, the robot arm is controlled based on the calculated displacement to correct the position of the end effector.

Here, step S3 will be described in more detail. The positioning marks are arranged at every 90 degrees on a concentric circle centered on the position where the end effector should act, and the camera is moved around the center axis of the end effector by 9 degrees.
Since the mark is rotated by 0 degrees to image the mark, if the mark is imaged at the same position in each field of view, the center axis of the position where the end effector should act and the center axis of the end effector coincide. That is easy to understand. The rotational displacement about the center axis is calculated from the displacement of the mark image position from the center of the field of view and the distance between the camera optical axis and the end effector center axis. FIG. 3 shows an example in which four marks are imaged. In the figure, the dashed cross marks indicate the center of the visual field, and 100-1 to 100-4 are marks 1-1 to 1 respectively.
In the field of view at the time of imaging of -4, the large and small circles indicated by broken lines indicate the boundaries of the side surfaces of the truncated cones of the respective marks that are shown when the positioning of the end effector was correct, and the large and small circles indicated by solid lines indicate the positioning of the end effector. It shows the boundary between the lower side and the upper side of the side of the truncated cone of each mark that is shown when it is shifted, and the upper and lower boundaries of this side are the two parts with different mark heights described in claim 2. Equivalent to.

A method of calculating the shift amount of the end effector when each mark is shifted from the center as shown by the large and small circles of the solid line in FIG. 3 will be described. As the displacement in the three-dimensional space, a translational displacement parallel to each axis of xyz and a rotational displacement about each axis of xyz must be considered. When this occurs, the position of the mark in the camera's field of view can be easily calculated by geometrical coordinate transformation.
However, conversely, the translation deviation amounts dx, dy,
It is generally not possible to analytically determine dz and the amounts of rotational deviation θx, θy, θz. Therefore, the deviation amounts dx, dy, d
Assuming z, θx, θy, and θz, the imaging position of each mark is calculated based on the assumption, and the sum of squares of the distance from the actual imaging position is used as an evaluation value V. d
y, dz, θx, θy, θz are obtained by using a mathematical method such as the steepest descent method.

This procedure will be described next. The origin of the coordinate system is a position that intersects the center plane of the camera lens on the center axis of the end effector. [1] First, it is assumed that the relative position between the work target and the end effector is as designed. That is, dx = 0, dy = 0,
It is assumed that dz = 0, θx = 0, θy = 0, and θz = 0. [2] Assumed dx, dy, dz, θx, θy, θz
, The spatial coordinates where each mark should be located are calculated from the design data. This is θ with respect to the designed spatial coordinates.
It is obtained by multiplying a so-called rotation matrix calculated from x, θy, and θz, and then adding translation vectors (dx, dy, dz). [3] From the calculated mark coordinates, the imaging position of the mark when the image is taken by the camera is calculated. This can be calculated from the equation of the lens of geometric optics. The center positions of the lower side and the upper side of the side surface of the truncated cone are generally different. [4] Next, as the evaluation value V, the square sum of the distance between the center position of the lower side and the upper side of the side of the truncated cone that is actually imaged and the imaging position calculated in [3] of the corresponding mark is used. Ask. If there is a shift, this value is a positive value other than zero. To determine the center positions of the lower side and the upper side of the side of the truncated cone, the side of the truncated cone is shown in black. [5] Next, the values of V when dx is changed by + Δx and when −x is changed are obtained in the same manner as in the steps [2] to [4], and are obtained by [4]. The dx that can obtain a value smaller than the value of V is adopted as a new dx. Δx is a small value. [6] dy, dz, θx, θy, θz are also [5]
A new value is adopted in the same manner as. [7] The steps [2] to [6] are repeated, and the evaluation value V
The calculation is terminated when a smaller value of is no longer obtained, and dx, dy, dz, θx, θy, θz at that time
Is the amount of deviation to be obtained. dx, dy, dz, θ
If x, θy, and θz are obtained, the position of the robot arm can be controlled according to the values to correct the position of the end effector. It should be noted that three or more marks are used because the unknowns are dx, dy,
This is because there are six values of dz, θx, θy, and θz, and only two values of the position xy can be obtained by observing one mark.

In the above description of the steps [1] to [7], it has been explained that both data on the upper side and the lower side of the truncated cone are used, but only one of them is used as dx, dy, dz, θ.
x, θy, and θz can be obtained. However, as described above, if the center positions of both the upper side and the lower side of the truncated cone are obtained and calculated from one image data, the amount of information is large, so the number of repetitions of steps [2] to [6] is reduced. There is an effect that can be reduced.

In the above description, a case where a truncated cone is used as a mark has been described. However, this may be a truncated pyramid, or may be a truncated cone or a truncated pyramid. Further, it does not have to have a three-dimensional structure.

[0015]

As described above, according to the present invention, a plurality of marks arranged on the same circumference centering on the position where the end effector is to act are provided at the tip of the robot arm integrated with the end effector. With the television camera attached to the axis, the robot arm axis is rotated by a predetermined amount and images are sequentially taken. Based on the position where the mark appears in the field of view of the television camera, the relative attitude correction amount of the end effector is calculated and the relative position of the robot arm is calculated. Since the posture is corrected, there is no need to make complicated marks as in the past, it is simple, the positioning accuracy of the center axis of the end effector is good, and multiple marks that are far apart from the rotational deviation around the center axis Therefore, there is an effect that accurate positioning can be performed as a result.

[Brief description of the drawings]

FIG. 1 is an explanatory diagram showing a robot arm relative posture correction method according to an embodiment of the present invention.

FIG. 2 is a flowchart illustrating a robot arm relative posture correction method according to an embodiment of the present invention.

FIG. 3 is an explanatory diagram showing a robot arm relative posture correction method according to one embodiment of the present invention.

FIG. 4 is a sectional view showing a positioning mark used in a conventional robot arm relative posture correction method.

FIG. 5 is a front view showing a positioning mark used in a conventional robot arm relative posture correction method.

FIG. 6 is an explanatory diagram illustrating a conventional robot arm relative posture correction method.

FIG. 7 is an explanatory diagram illustrating an occurrence of an error in a conventional robot arm relative posture correction method.

FIG. 8 is an explanatory diagram illustrating an occurrence of an error in a conventional robot arm relative posture correction method.

[Explanation of symbols]

 1-1 Positioning mark 1-2 Positioning mark 1-3 Positioning mark 1-4 Positioning mark 10 Camera 20 End effector 30 Robot arm 35 Camera mounting arm 40 End effector action position

──────────────────────────────────────────────────続 き Continuation of the front page (72) Inventor Miwako Hirooka 8-1-1, Tsukaguchi Honcho, Amagasaki City Mitsubishi Electric Corporation Industrial System Research Laboratory (56) References JP-A-1-193902 (JP, A) JP-A-62-95403 (JP, A) JP-A-2-264804 (JP, A) JP-A-64-91005 (JP, A) JP-A-2-300608 (JP, A) JP-A-1-320431 (JP JP, A) (58) Field surveyed (Int. Cl. ⁶ , DB name) G05B 19/19 B25J 9/10 B25J 13/08 B25J 19/04 G01B 11/00

Claims (1)

(57) [Claims] 1. In a work object of a robot, on the same circumference around a position where an end effector should act,
Three or more marks are arranged, in the end effector and a TV camera attached to the robot arm shaft are integrated, the robot arm shaft is rotated by a predetermined amount by sequentially imaging the marks, captured Measure the position of the mark in the same field of view of the TV camera,
The relative posture correction amount of the end effector is calculated based on the measurement result, and the end effector is calculated based on the relative posture correction amount.
A robot arm relative posture correction method for correcting the posture of the effector .