JPH04255003A - Off line teaching system - Google Patents

Abstract

PURPOSE:To form a highly precise operation program by off line teaching. CONSTITUTION:An arithmetic means 12 finds out a conversion matrix A based upon the arrangement data 11 of a model. An arithmetic means 13 finds out a conversion matrix B based upon arrangement data obtained when a robot touches up a practical work. An arithmetic means 14 finds out a conversion matrix C for converting the matrix A to the matrix B. An arrangement data correcting means 15 corrects the arrangement data 11 of the model by multiplying the arrangement data by the matrix C. The correction allows the arrangement of the model 1 to coincide with the practical work arrangement. The operation program of the robot is formed by using the corrected arrangement data.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an offline teaching method used for teaching a robot offline, and more particularly to an offline teaching method for accurately operating a robot in an actual working environment.

0002

[Prior Art] In robots used in automobile welding lines, etc., in order to shorten the start-up time of the line,
Operation programs created through offline teaching have become essential. This operation program is created based on a computer model within the offline programming system and transferred to the robot control device. The robot performs work on the line based on its operation program.

0003

[Problems to be Solved by the Invention] However, in offline programming systems, there are placement errors in robot and workpiece models, and the computer model and the actual model are usually quite different. For this reason, operation programs created based on computer models that include placement errors, etc.
The accuracy was not very high, and even if used as is, it would not be possible to make the robot perform the specified movements. Therefore, in order to make the robot perform a specified action,
Considerable modifications had to be made to the operating program. The present invention has been made in view of these points, and it is an object of the present invention to provide an offline teaching method that allows highly accurate operation programs to be created.

0004

[Means for Solving the Problems] In order to solve the above problems, the present invention uses an offline teaching method in which a robot is made to perform a predetermined motion based on a program created with an offline programming system. a first calculation means for calculating a first transformation matrix representing a coordinate system of a plane formed by the three points on the model, based on data obtained by specifying three points on a graphic display; Based on the data obtained by touching up the actual three points on the object corresponding to the three points on the object using a robot, the coordinate system of the plane formed by the three points on the object is calculated. a second calculation means for calculating a second transformation matrix; a third calculation means for calculating a third transformation matrix for converting the first transformation matrix into the second transformation matrix; , a placement data correction means for correcting placement data of the model, and an operation program creation means for performing off-line teaching of the robot and creating an operation program for the robot based on the corrected placement data of the model. An offline teaching method is provided.

[0005]

[Operation] Specify three points on the model of the object on the graphic display. A first transformation matrix is determined based on the data obtained by the specification. This first transformation matrix represents a coordinate system of a plane formed by three points on the model. Next, a robot is used to touch up three actual points on the object that correspond to the three points on the model. A second transformation matrix is determined based on the data obtained by the touch-up. This second transformation matrix is the actual 3
Represents the coordinate system of the plane formed by the points. Furthermore, a third transformation matrix for transforming the first transformation matrix into a second transformation matrix is determined. The placement data of the model is corrected using this third transformation matrix. Therefore, the arrangement of the model and the actual arrangement of the object match. Based on this corrected placement data, off-line teaching of the robot is performed and an operation program for the robot is created. Therefore, a highly accurate operation program can be obtained.

[0006]

DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. FIG. 2 is an overall configuration diagram for implementing the present invention. The offline program device 10 creates a robot operation program offline. For example, a workstation is used as the offline program device 10. In the offline programming device (workstation) 10, a model 51 of the workpiece 50 is displayed on the graphic display 10A. That model 5
1, four points P1, P2, P3, and P4 are picked up, and the arrangement data 11 (FIG. 1) of the four points P1 to P4 is set in the offline programming device 10. This pickup is performed using the mouse 10B. 4 points P1-P
4 are not on the same plane. Robot control device 3
0 has a microprocessor configuration, which drives the servo motor of the robot 40 to operate the robot 40 based on an operation program. A touch-up sensor 42 is provided at the tip of the arm 41 of the robot 40. A proximity switch, a capacitance switch, or the like is used as the sensor 42. The robot 40 uses the sensor 42 to detect four points P1A on the workpiece 50.
, P2A, P3A, and P4A, and touch those four points P1
The arrangement data 31 (FIG. 1) of A to P4A is set in the robot control device 30. These four points P1A to P4A correspond to the four points P1 to P4 on the model 51. The offline program device 10 stores arrangement data 11 and 31.
Based on this, the arrangement data of the model 51 is corrected (calibrated), and teaching is performed based on the corrected arrangement data. The details are described below.

FIG. 1 is a block diagram of the offline teaching method of the present invention. Offline programming device 10
The calculation means 12 calculates the transformation matrix A based on the arrangement data of three points, for example, three points P1, P2, and P3 out of the arrangement data 11 of the four points P1 to P4 on the model 51. This transformation matrix A represents a coordinate system of a plane formed by three points P1, P2, and P3. Here, the reason why the three points are limited from the arrangement data 11 is to make the coordinate system an orthogonal coordinate system, and this orthogonal coordinate system maintains consistency between each model. The transformation matrix A is obtained as follows. First, consider the rotating element. The vector from point P1 to point P2 is normal vector n, the normal vector to the plane formed by three points P1, P2, and P3 is approach vector a, and the vector perpendicular to normal vector n and approach vector a is orientation vector o. shall be. Next, the translation element is a
1 (position vector of point P1). At this time, the transformation matrix A is expressed by the following equation (1). However, normal vector n = (nx ny
nz) Orient vector o=(ox oy oz
) approach vector a = (ax ay az)

[
[0008] The calculation means 13 calculates the location data 31 of the four points P1A to P4A set in the robot control device 30.
3 points P1A, P2A, P3A (3 points P1 of model 51,
A transformation matrix B is obtained based on the arrangement data of points corresponding to P2 and P3. Transformation matrix B has three points P1A, P2A
, P3A, and is obtained in the same manner as the transformation matrix A described above. The calculation means 14 calculates the transformation matrix C using the following equation (2). Transformation matrix C transforms matrix A into matrix B. C=invB×A──────(2) However, invB is B's inverse matrix placement data correction means 15 corrects the placement data of the model 51 by multiplying the placement data of the model 51 by the transformation matrix C. do. This correction allows the arrangement of the model 51 to match the arrangement of the actual workpiece 40. The operation program creation means 16 performs off-line teaching using this corrected arrangement data, and the operation program for the robot 40 is created by this teaching. After the operation program has been verified through simulation, it is transferred to the robot control device 30. The robot 40 operates and performs work based on the transferred operation program 32. In this way, the arrangement data of the model 51 is corrected using the transformation matrix C, so that the arrangement of the model 51 matches the arrangement of the actual workpiece 50. Therefore, an operation program created based on this corrected placement data is not affected by errors in the placement data. Therefore, a highly accurate operation program can be obtained.

FIG. 3 is a block diagram showing a second embodiment of the offline teaching method of the present invention. The difference from the first embodiment is that the operating program is configured to be corrected. In the figure, the calculation means 17 of the offline program device 10 operates at four points P1 to P1 of the arrangement data 11.
A transformation matrix D is determined based on P4. This transformation matrix D represents a coordinate system of a plane formed by four points P1, P2, P3, and P4. Here, from the four points of the arrangement data 11, the transformation matrix D
The reason for calculating this is to make the coordinate system a distorted coordinate system (a coordinate system that is not an orthonormal coordinate system) to correspond to the distortion of the surface formed by the four points P1 to P4, thereby further improving the accuracy of the motion program. This is to make it happen. The calculation means 18 calculates the four points P1A of the arrangement data 31 set in the robot control device 30.
- Find the transformation matrix E based on P4A. transformation matrix E
represents a coordinate system of a plane formed by four points P1A, P2A, P3A, and P4A. As mentioned above, this coordinate system is 4
A coordinate system is distorted corresponding to the plane formed by points P1A to P4A. The calculation means 19 obtains a transformation matrix F. This transformation matrix F transforms the matrix D into the matrix E, and is obtained in the same manner as the transformation matrix C described above. The motion program correction means 20 uses this conversion matrix F to correct (calibrate) the motion program, and the corrected motion program is transferred to the robot control device 30. In this way, in the second embodiment, the motion program is corrected based on the coordinate system (distorted coordinate system) obtained from the arrangement data of four points. Therefore, the accuracy of the first motion program created based on the orthogonal coordinate system is further improved, and the robot 40 can perform a predetermined motion even if used as is without any modification. FIG. 4 is a block diagram showing a third embodiment of the offline teaching method of the present invention. The difference from the second embodiment is that the motion program is corrected in the robot control device 30. That is, the calculation means 17, 18
, 19, and operation program correction means 20 are provided in the robot control device 30.

0010

As described above, in the present invention, the arrangement data of the model is corrected so that the arrangement of the model matches the arrangement of the actual object. Therefore, the operation program created based on this corrected placement data is
Not affected by errors in placement data. therefore,
A highly accurate motion program can be obtained, and the robot can be used as is to perform predetermined motions without any modification.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of an offline teaching method of the present invention.

FIG. 2 is an overall configuration diagram for implementing the present invention.

FIG. 3 is a block diagram showing a second embodiment of the offline teaching method of the present invention.

FIG. 4 is a block diagram showing a third embodiment of the offline teaching method of the present invention.

[Explanation of symbols]

10 offline program device 11 arrangement data 12, 13, 14, 17, 18, 19 calculation means 16
Motion program creation means 20 Motion program correction means 30 Robot control device 31 Touch-up arrangement data 32 Motion program 40 Robot 50 Work 51 Model

Claims (5)

[Claims] Claim 1: In an offline teaching method in which a robot performs a predetermined action based on a program created with an offline programming system, data obtained by specifying three points on a model of an object on a graphic display is used. a first calculation means for calculating a first transformation matrix representing a coordinate system of a plane formed by the three points on the model based on the three points on the model; and three actual points on the object corresponding to the three points on the model. a second calculation means for calculating a second transformation matrix representing a coordinate system of a plane formed by three points on the object, based on data obtained by touch-up using a robot; a third calculation means for calculating a third transformation matrix that converts the first transformation matrix into the second transformation matrix; and a layout data correction means that corrects the layout data of the model using the third transformation matrix; An offline teaching method comprising: an operation program creation unit that performs offline teaching of the robot and creates an operation program for the robot based on corrected placement data of the model. 2. Based on data obtained by indicating four points on the model that are not on the same plane on a graphic display, a fourth point representing a coordinate system of a surface formed by the four points on the model is determined. a fourth calculation means for calculating a transformation matrix; and a fourth calculation means that calculates a transformation matrix for the object based on data obtained by touch-up using a robot to touch up four actual points on the object corresponding to the four points on the model. a fifth calculation means for calculating a fifth transformation matrix representing a coordinate system of a surface formed by the four points above; and a sixth calculation means for calculating a sixth transformation matrix for converting the fourth transformation matrix into the fifth transformation matrix. 2. The offline teaching method according to claim 1, further comprising a sixth calculation means and a motion program correction means for correcting the motion program using the sixth transformation matrix. 3. The offline teaching method according to claim 2, wherein the fourth calculation means, the fifth calculation means, the sixth calculation means, and the operation program correction means are provided in the offline programming system. . 4. The offline teaching method according to claim 2, wherein the fourth calculation means, the fifth calculation means, the sixth calculation means, and the motion program correction means are provided in a robot control device. 5. Claim 2, wherein the three points on the model and three out of four points on the model are common.
Offline teaching method described.