JPS63105893A - Automatic teaching method of robot - Google Patents

Abstract

(57) [Abstract] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention relates to a method for teaching the position of a robot, and is particularly suitable for simplifying and manpower-saving the troublesome teaching work that requires a lot of man-hours and time. Concerning automatic teaching methods for robots.

[Conventional technology]

In conventional robot teaching methods, when a movement button on the teaching box is pressed, the robot moves in the direction indicated by the button, such as the X-axis direction, for as long as the button is pressed, and it is difficult to check whether the robot has reached a predetermined position or posture. The operator visually checks and positions the parts. In addition, when using a visual detector, etc., a method is used in which the shape of the object to be taught is recognized and the teaching point of the motion path is determined based on the characteristics of that shape, such as two temporary joining lines in the case of a welded object. ing.

Such conventional teaching methods require a lot of teaching man-hours and time, which puts a heavy burden on the operator. No consideration was given to automatic teaching. Regarding conventional robot teaching methods, for example, Japanese Patent Application Laid-open Nos. 59-177610 and 1983-
It is described in Publication No. 39207 and the like.

[Problem that the invention seeks to solve]

The above-mentioned conventional technology mainly involves recognizing the object to be taught and detecting the characteristic points of its shape, such as the joining line of the welded product, and determining it as the teaching point. This method requires a considerable amount of time, up to 2.5 minutes, and has a problem in that teaching precision is limited, making it difficult to teach complex assembly tasks.

The purpose of the present invention is to provide a simplified automatic robot that can automatically teach positioning points and postures based on the information provided with marks as teaching aid information for determining gripping positions, postures, etc. on parts to be taught. To provide a teaching method.

[Means for solving problems]

The above purpose is to grasp the gripping position, direction, and direction of the teaching point on the teaching target part.
A teaching mark is provided that has the meaning of teaching auxiliary information for determining posture, etc., and the teaching mark is detected by an image sensor and the teaching auxiliary information is captured, and the position of the teaching mark is measured and the teaching point is determined. This is achieved by an automatic robot teaching method that calculates and teaches the position, orientation, etc. of the robot.

[Effect]

In the robot automatic teaching method described above, the teaching mark of the teaching target part is captured by an image sensor, and the type of teaching point (
In addition to recognizing the gripping point, insertion point, etc.), the position of the recognized teaching mark is measured to determine the three-dimensional coordinate value in the robot coordinate system, and the robot hand is By calculating the positioning position, posture, etc. of the machine, and inputting information that cannot be specified from the mark information (grip depth of the hand, etc.) as part attachment information, the teaching point information can be obtained accurately and quickly by the computer. be able to.

〔Example〕

An embodiment of the present invention will be described below with reference to FIGS. 1 to 8.

FIG. 1 is a block diagram showing the overall configuration of an embodiment of the automatic robot teaching method according to the present invention. In Figure 1,
1 is an overall TV camera for inputting the entire image, 2 is a hand TV camera attached to the robot's hand, 3 is an image processing device, 4 is an arithmetic processing device, 5 is a memory for storing mark information, 6
is a memory for storing teaching information, 7 is a distance measuring sensor for position detection, 8 is a teaching target part attached by pasting a teaching mark, etc.
9 is a teaching box.

FIG. 2 is an explanatory diagram of the storage format of the mark information shown in FIG. 1. In Figure 2, 10 is the mark information serial number, 1
1 is the mark name, 12 is the meaning of the mark, and 13 is the feature amount of the mark (standard area 1 perimeter length, etc.). Further, FIG. 3 is an explanatory diagram of the storage format of the parts information shown in FIG. 1. In FIG. 3, 14 is a part information serial number, 15 is a part name, and 16 is teaching attached information (grip depth d, etc.) of the part. FIG. 4 is an explanatory diagram of an example of the teaching mark shown in FIG. 1 (FIG. 2). In FIG. 4, 17 is an example of the teaching mark of each mark number 1 to 6, and 18 is an example of the meaning of the mark.

5(a), (bl is an explanatory diagram of an example of the teaching mark on the flat grip part of FIG. 1 (FIG. 4), FIG. Figure (bl is a perspective view showing the relationship between parts and robot hands. Figure 5 (a)
, (In bl, 19, 20, 21 have the meanings of the 1st, 2nd, and 3rd teaching auxiliary information of the flat grip part with mark numbers 1, 2, and 3 attached to the teaching target part (Fig. 1) Of the teaching marks (FIG. 4), teaching marks 19 and 20 indicate the gripping directions a and b and the gripping position of the robot hand 22,
The teaching mark 21 is used to find the normal vector of the plane of the part 8 when making the hand surface A and the part 8 parallel. Note that the gripping depth d is stored as component attachment information (FIG. 3).

In the configuration shown in FIG. 1, a procedure for performing automatic teaching processing on the teaching target part 8 and obtaining teaching point information will be explained. First, by selecting and pressing the automatic teaching mode button on the teaching box 9, the manual operation/automatic teaching mode is entered (activation of automatic teaching). Next, the hand TV camera 2 is moved onto the first teaching mark (first teaching point) 19 of the part 8 based on the information taken in from the overall TV camera 1. Alternatively, the user moves the hand lever 2 manually to bring the first teaching mark 19 into the field of view (movement of the image sensor).

Here, the arithmetic processing device 4 issues a command to take in the image data of the part 8 to the image processing device 3, and the image processing device 3
Image information is taken in from the V camera 2. Next, the arithmetic processing device 4 issues a command to inquire whether the first teaching mark 19 attached to the component 8 is within the field of view, and the image processing device 3 determines that the teaching mark 19 is present if it is within the field of view. ” is not returned, a response of “None” is returned. When the arithmetic processing unit 4 is "absent", it inquires about the presence or absence of the next second teaching mark, but if there is a first teaching mark 19, its position, area, and orientation information (grasping point of the type of teaching point, (insertion point, etc.) and imports that information. Next, perform the second process using the same processing procedure as above. The information related to the third teaching mark 20.21 is taken in (teaching mark importing).

FIG. 6 is an explanatory diagram of position/orientation information within the image processing device 3 of FIG. 1. In FIG. 6, 23 is the field of view of the hand TV camera 2, 24 is the origin of the camera field of view, 25 is the origin of the xy coordinates within the field of view, 26 is the principal axis of inertia of the shape of the teaching mark,
27 is the center of gravity of the teaching mark.

Normally, the positional information related to the teaching mark of the arithmetic processing device 3 is obtained as the xy coordinate values of the center of gravity (position) 27 of the xy reference coordinate system defined within the field of view 23 of the hand TV camera 2, and is not particularly specified. If not, the origin of the camera field of view 24
One of the posture information is obtained as the principal axis of inertia 26 of the mark shape, for example. Note that this type of image processing method is a function of the normal image processing device 3;
For details, for example, see the key point "Small image processing device suitable for position/shape measurement" Hitachi Review Voβ, 67, 9th floor (198
Although it is also described in literature such as PP, 67-70 (September 1982) and Takano "Shape Pattern Recognition Technology" Information Research Group (October 1982), it is omitted here. Next, the arithmetic processing unit 4 then processes the teaching marks 19, 20 . . . recognized as described above.
The position of the mark 21 is measured to obtain three-dimensional coordinate values in the robot coordinate system.

FIG. 7 is a block diagram of an embodiment of the mark position measuring method shown in FIG. 1. In Fig. 7, 28 is an actuator (motor), 29 is a servo circuit, and 30 is a servo voltage value (
31 is the current position (No. 13), 32 is the servo voltage (signal), 33 is the limit signal, 34 is the proximity switch (
(corresponds to the distance measuring sensor 7).

Here, the arithmetic processing unit 4 moves the tip 22 of the previously determined center of gravity position 27 of the teaching mark. This requires a certain amount of time (servo period) for the servo circuit 29.
This is done by outputting a servo voltage value of 30 every t, and since the servo circumference 'Mj is determined at this time, if a command is given to a position farther from the current position, the speed will be faster, and if a command is given to a position closer to the current position, the speed will be slower. This servo voltage value 30 indicates the rotation speed of the actuator (motor) 28, so in order to obtain this servo voltage value 30, it is necessary to know the rotation angle of each joint of the robot to move to a certain position. The methods of
PP, 63-70, but the details are omitted here. In this way, the arithmetic processing device 4
The proximity sensor 34 attached to the hand 22 of the robot slowly moves toward the center of gravity 27, and at the same time constantly monitors the limit signal 33 from the proximity sensor 34. When the proximity switch 34 touches the center of gravity 27, the limit signal 33 is input to the arithmetic processing unit 4, and at that point the arithmetic processing unit 4 sets the servo voltage value 30 to zero and stops the movement of the robot. Next, input the current position signal 31 of the actuator (motor) 28, which is usually obtained from an encoder attached to the motor shaft.
The center of gravity position of the teaching mark 270 determines the three-dimensional spatial coordinate value in the robot coordinate system. Note that the method of determining the coordinate values from the encoder values of this motor is also described in the above-mentioned computer roll section 9, but the details will be omitted.

In this way, the mark positions of the required number of teaching marks 19, 20, and 21 are measured (mark position measurement). Next, the arithmetic processing unit 4 calculates the positioning position and posture (teaching point) of the hand based on the measured value (coordinate value) of the previously determined mark position and the meaning of the recognized teaching mark (teaching point type).

FIG. 8 is an explanatory diagram of an example of position/posture information of the robot hand in FIG. 1. In FIG. 8, 35 is a hand position (vector) P, 36 is a hand posture (vector), and f137 is a hand posture (vector) g. Here, the arithmetic processing unit 4 calculates the gripping position, orientation, etc. of the component 8 based on the measured values (coordinate values) of the mark position obtained previously. For example, the center of gravity position 2 of the teaching marks 19, 20, and 21 (Fig. 5) found earlier.
The position vector in the robot coordinate system such as 7 is P
I4, PZO.

As pz+, as shown in FIG. 8, the grip center of the hand 22 is the hand position 35 (hand position vector P), the direction of the hand 22 is the hand posture 36 (hand posture vector f), and the direction perpendicular to the direction of the hand 22 is the hand posture. If the position and posture of the hand are determined as 37 (hand posture vector g), the position and posture of the grip are calculated by the following equation.

f=(p2゜−PI9) X (Pz+PI9)
(11g=Pz+PI9
(21P=(P+w”Pz+) d-f
f3) The symbol attached to the vector here indicates unitization of the vector (calculation of teaching points).

Here, the gripping depth d used in equation (3) above may be determined in advance as part attachment information 16 (Fig. 3), or
Each time, it is determined by inputting it from the teaching box 9 and stored as part information (input of attached information). In this way, the information on the hand position 35 and hand posture 36, 37 obtained by equations (1) to (3) is stored in the teaching information memory 6 as teaching point information.

Note that the measurement of the mark position by the proximity sensor 34 shown in FIG. 7 may be performed directly by using an ultrasonic sensor or the like.

Furthermore, the teaching mark shown in FIG. 4 above is just an example, and teaching marks with meaning as necessary teaching auxiliary information can be provided depending on the shape of the object to be taught, etc., so that the grasping position of the teaching point can be Not only the posture but also the gripping method and assembly/
The insertion method, etc. may also be requested.

〔Effect of the invention〕

According to the present invention, by attaching the teaching 5 mark as teaching auxiliary information to the teaching target part of the robot, it becomes possible to automatically capture teaching points accurately and easily.
This is effective in reducing teaching man-hours and time as well as improving teaching accuracy.

[Brief explanation of the drawing]

Fig. 1 is a block diagram of the overall configuration showing an embodiment of the robot automatic teaching method according to the present invention, Fig. 2 is a storage format diagram of the mark information shown in Fig. 1, and Fig. 3 is a storage format of the part information shown in Fig. 1. 4 is an explanatory diagram of the teaching mark example shown in FIG. 1, FIG. 5 Tag, and (b) is a perspective view showing the relationship between the top view of the teaching mark example of the flat grip part of FIG. 1 and the hand. , FIG. 6 is an explanatory diagram of the position/orientation information in the image processing device shown in FIG. 1, FIG. 7 is a block diagram of an example of mark position measurement shown in FIG. 1, and FIG. FIG. 3 is an explanatory diagram of an example of position/posture information of a hand. 1... General camera, 2... Hand camera, 3...
Image processing device, 4... Arithmetic processing device, 5... Memory for storing mark information, 6... Memory for storing teaching information, 8
...Teaching target part, 9...Teaching box,
12...Meaning of marks, 16...Parts attachment information, 1
7...Teaching mark, 18...Meaning of mark, 19.
20.21...Mark, 22...Stooge, 27...
Center of gravity, 26... Principal axis of inertia, 35... Hand position, 36
．． 37... Hand posture. Agent Patent Attorney Tadashi Akimoto No. 1 Figure 4 Figure 5 (a) No. 8 (b) Figure 7 Figure 8 Figure 35 Child

Claims (1)

[Claims] 1. Attach a teaching mark with meaning as teaching auxiliary information to obtain the grasping position, posture, etc. of the teaching point on the part to be taught by the robot, detect the teaching mark, recognize the meaning, and use the teaching mark. An automatic robot teaching method that calculates and teaches the grasping position, posture, etc. of the robot's teaching point based on that information by detecting the position of the robot.