JPH10207524A - Method for automatic off-line teaching of sensing pattern and method for simulating sensing operation - Google Patents

Abstract

PROBLEM TO BE SOLVED: To generally and automatically execute the preparation of teaching data or the simulation of operation in various sensing formats of respective robot manufacturers. SOLUTION: A specific sensing pattern code is selected by a key and a sensing operation pattern corresponding to the selected sensing pattern code is selected by using a sensing operation pattern data base for defining the shapes of members and plural kinds of sensing operation patterns different in respective robot manufacturers and a sensing pattern code data base 17 for defining plural kinds of sensing pattern codes 0 to 10,... corresponding to respective sensing operation patterns stored in the sensing operation pattern data base by the use of keys such as the shapes of members and the restricting conditions of respective robots to prepare teaching data from the sensing operation pattern, to allow a robot to execute sensing operation or execute the simulation of sensing operation.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to teaching of a sensing operation (position of a robot) necessary for performing touch sensing (contact sensing) or non-contact sensing for various robots such as welding and painting. And teaching methods relating to posture, operation method, welding method, etc.) and a method for automatically performing operation simulation offline.

[0002]

2. Description of the Related Art Conventional techniques will be described by taking a welding robot as an example.

When a robot is used for welding, it is necessary to hold the tip of the welding wire correctly at the welding line and perform welding. Data required for welding (the robot position at each point,
Teaching of the robot to welding conditions, operating conditions, etc.) is performed. However, if the member is a gas cut material and its accuracy is poor, or there is an error in the installation of the member with respect to the robot, and there is an error in the positional relationship between the robot and each member, welding can be performed using only the teaching data. It cannot be done.

Therefore, an electric potential is applied between the welding wire and the member, and the robot touches (contacts) the tip of the welding wire with the member, thereby detecting the ON of the short-circuit signal. The mainstream method is to use a "touch sensing function" that corrects the position of a member to correct an error of a member (poor accuracy or installation error) and to perform welding along a welding line.

[0005] On the other hand, as teaching of operating conditions and welding conditions for a welding robot, conventionally, while an operator actually moves a robot in a work place, teaching of a robot position at each point, welding conditions and operating conditions at that point are performed. Was teaching.

Therefore, the time during which the worker is teaching is a time during which the robot cannot actually perform welding, and there is a problem that the operating rate of the robot is reduced accordingly.

[0007] Accordingly, an offline automatic teaching system has been developed in which teaching data is created in advance in an office.

[0008] However, in the current off-line automatic teaching system, since the method of moving the robot to a member necessary for touch sensing and the method of instructing the welding robot differ depending on the manufacturer of the welding robot, it is difficult to make a general-purpose program. Teaching of touch sensing and motion simulation could not be performed.

For example, in touch sensing for recognizing the start end of a fillet member, there are the following methods of moving the robot. However, since the specific operation method and order are individually determined by the welding robot manufacturer, There are various ways.

In order to sense the corner, the welding torch is moved in the front direction, and the position of the standing abdominal surface is recognized by touch sensing with the wire tip of the welding torch or the abdominal surface of the welding torch. Further, the welding torch is moved downward, and the position of the bottom surface of the member surface is recognized by touch sensing with the wire tip of the welding torch or the abdominal surface of the welding torch.

Further, in order to recognize the end of the member,
The welding torch is moved sideways and horizontally, and the position of the standing member side is recognized by touch sensing on the side surface of the welding torch or the abdominal surface of the welding torch. Further, the welding torch is moved downward and left and right, and the position of the bottom surface of the member surface is recognized by touch sensing with the side surface of the welding torch or the abdominal surface of the welding torch.

Further, assuming that the members are not mounted vertically, three points on the abdominal surface on the side of the standing member and three points on the bottom surface of the member surface are touch-sensitive to determine the position of the surface of each member. Recognize the tilt.

As described above, since the specific operation method and order of touch sensing differ for each welding robot maker, when teaching or operation simulation is performed in off-line automatic teaching, control of the touch sensing portion is performed by a program. At present it is not incorporated.

Therefore, in the current operation simulator of the offline automatic teaching, a manual operation is performed in which the operator actually moves the tip of the welding robot (the tip of the welding wire) while watching the monitor screen and teaches the operation of touching the member. Was needed.

[0015] For this reason, when the number of target members of the touch sensing increases, the work becomes complicated, and as a result, the operation simulator cannot be used for an object handling a large number of members.

The above is a description of touch sensing in off-line automatic teaching of a welding robot. However, even when the robot is used for various operations such as painting, a non-contact detection of the positional relationship between the robot and members without contact is also performed. A similar problem exists in the case of contact-type sensing.

[0017]

SUMMARY OF THE INVENTION In view of the above-mentioned problems of the prior art, the present invention automatically and universally generates teaching data or simulates operation for touch sensing or non-contact sensing of each manufacturer. It is an object to provide a method that can be performed.

[0018]

According to the present invention, there is provided a sensing pattern offline automatic teaching method according to the present invention, comprising: a sensing operation pattern database for defining a plurality of types of sensing operation patterns different for each member shape and each robot maker; Using a sensing pattern code database that defines a plurality of types of sensing pattern codes corresponding to each sensing operation pattern in the same database using keys such as member shapes and robot constraints, and using the keys to read from the sensing pattern code database. Select a specific sensing pattern code, select a sensing operation pattern corresponding to the specific sensing pattern code from the sensing operation pattern database, and select a sensing operation pattern according to the selected sensing operation pattern. It is characterized in that automatically create teaching data offline.

In the method for automatically teaching a sensing pattern off-line according to the present invention, the sensing is a touch sensing, or the sensing is a sensor using a laser beam, a sensor using a camera, or It is characterized by non-contact sensing using sensors using electricity such as eddy current sensors and capacitance sensors, or the sensing is performed by a welding robot to recognize the beginning and end of a member, recognize the groove shape, etc. There is a feature.

Further, a sensing operation simulation method according to the present invention for solving the above-mentioned problems is characterized in that a sensing operation of a robot is simulated on a graphic display using teaching data created by any of the above methods. It is.

[0021]

Embodiments of the present invention will be described below. In the present embodiment, in the off-line automatic teaching system, in the case of wire contact type touch sensing in which there are several types provided by each welding robot maker, when this operation is simulated or when teaching data is created, this operation is performed. It is an offline automatic teaching system that has a file in which a database of teaching patterns and teaching patterns is created and can be used for welding robots provided by any manufacturer for general purposes. As the database, a touch sensing operation pattern database that defines a plurality of types of sensing operation patterns that are different for each member robot and a welding robot maker, and a plurality of types of sensing pattern codes corresponding to each sensing operation pattern from the database, It consists of a touch sensing pattern code database defined by keys such as shape and robot constraints.

By having such a database, touch sensing can be controlled in a general-purpose manner. In addition, based on the teaching data (see reference numeral 12 in FIG. 1) that has been automatically taught, welding can be actually performed on site by a welding robot, or operation simulation can be performed on a screen.

Hereinafter, an embodiment of the present invention will be described with reference to the drawings, taking a wire contact type touch sensing of a welding robot as an example. Here, FIG. 1 is a diagram illustrating the entire system configuration, FIG. 2 is a diagram illustrating an outline of the processing flow, and FIG. 3 is a diagram illustrating the touch sensing pattern code database 17.
Diagram explaining the example of touch sensing pattern code inside,
FIG. 4 is a diagram illustrating an example of a touch sensing operation pattern in the touch sensing operation pattern database, and FIG. 5 is a diagram illustrating an example of the touch sensing operation pattern 18.

As shown in the system configuration of FIG. 1, in the present embodiment, the CAD data 11 for the welding member is first read, and the teaching data 12 is converted into the three-dimensional off-line automatic teaching system 13 in accordance with the shape of the welding member and other instructions. Create with.

At this time, a three-dimensional operation simulation 14 of the welding robot is performed on the graphic display based on the teaching data 12.

At this time, if the welding robot interferes with the welding member, the interference check list 15 is displayed or printed out.

At this time, the teaching database 16 controls how the sensing operation is performed, and various teaching operations are created by the teaching database 16.

In the present embodiment, the teaching database 1
6 stores a touch sensing pattern code database 17 in which various touch sensing pattern codes 0 to 10 shown in FIG. 3 are stored, and various touch sensing operation patterns such as a touch sensing operation pattern 18 illustrated in FIG. It consists of a touch sensing operation pattern database.

In the touch sensing operation pattern database, various touch sensing operation patterns 18 are different depending on the member shape and the welding robot manufacturer. In the touch sensing pattern code database 17, each of the touch sensing pattern codes 0 to 10... Is defined by a key such as a shape of a member or a constraint condition of a robot, and the address of each sensing operation pattern 18 in the touch sensing operation pattern database, that is, Corresponding to the touch sensing pattern table addresses at0 to at10.

Next, an example of a procedure for producing the sensing operation will be described with reference to FIGS. 2, 3 and 4. FIG.

First, assuming that there are a plurality of welding lines,
The welding number n is initially set (n = 1), and the following processing is performed for each welding number n (see steps S1 and S2 in FIG. 2).

Although details will be described later, the design CAD data 11
Are determined as key data in steps S3 to S9 in FIG. 2 and touch sensing pattern codes 0 to 10... Are determined in step S10 based on the touch sensing pattern code database 17 in FIG.

As the key data, in the example of FIG. 3, the shape of the member, the presence or absence of the adjacency, and what surface is in contact? , Welding posture,
The detection of the end of the member, the detection of the inclination of the member surface, the detection of the detection of the inclination of the end of the member, and the like.

In order to obtain a value for each key data, subroutines sub0 to sub0 as shown in FIGS.
6 is called to perform processing. By calling the subroutine in this way, if the name of the subroutine to be called is changed, an output can be obtained by changing the value of any intended key data. Also, regarding the combination of key data, by changing the number of subroutine calls,
Any combination can be set.

A specific example will be described below with reference to FIGS.

First, a subroutine sub0 is called in step S3 of FIG. 2 to automatically recognize a member shape based on the design CAD data 11. Is the touch sensing pattern code database 17 of FIG. 3 a rib for fillet welding a member vertically? Is it a box-shaped thing? Is it a pipe member? This is an example of recognizing a rough difference in shape.

Next, the subroutine sub1 is called in step S4 of FIG.
The presence or absence of the neighbor is automatically determined. When sensing near the end of a member, interference occurs if there is another member near the end at the time of sensing. Therefore, the degree of approach of another member to the member end is checked. In the example of FIG. 3, the degree of approach is adjacent = “no”, adjacent = “present”, and adjacent = “narrow”
Are classified into three types. "None" indicates the degree of adjacency that can be sensed in a completely free state, "Yes" indicates the degree of adjacency that requires sensing while considering interference, and "Narrow" indicates that the distance to other members is small. At the end, it indicates the degree of adjacency that cannot be sensed.

In step S5 of FIG.
b2 is called, and based on the design CAD data 11, what surface is in contact? Is automatically determined. This is because the sensing method changes depending on how many surfaces the end of the member contacts. The member and the rib material have "two faces", and the box material has "three faces".

In step S6 of FIG.
b3 is called, and the welding attitude is automatically obtained based on the design CAD data 11. In the example of FIG. 3, there are a horizontal fillet, a downward groove, a lateral groove, a vertical fillet, an upward fillet, an upward groove, and so on.

In step S7 of FIG.
b4 is called, and based on the design CAD data 11, the presence or absence of the detection of the member end, that is, whether or not the member end is detected is automatically determined. This is due to the mounting accuracy of members, processing accuracy,
Determined by manufacturing accuracy.

In step S8 of FIG.
b5 is called, and based on the design CAD data 11, the presence / absence of the detection of the inclination of the member surface, that is, whether or not the inclination of the member surface is detected is automatically determined. This is also determined based on the mounting accuracy of the member, the processing accuracy, the manufacturing accuracy, and the like, similarly to the detection of the end of the member.

In step S9 of FIG.
b6 is called, and based on the design CAD data 11, it is automatically determined whether or not the end of the member is detected, that is, whether or not the end of the member is detected. This is also the same as the presence / absence of detection of the member surface tilt, as well as the member mounting accuracy, processing accuracy,
Determined by manufacturing accuracy.

Next, in step S10 of FIG. 2, the touch sensing pattern code is automatically determined. That is, based on the values of the key data obtained in steps S3 to S9, FIG.
From the touch sensing pattern code database 17 corresponding to the key data are selected.

Next, in step S11 of FIG. 2, a touch sensing operation pattern corresponding to the selected touch sensing pattern code is further selected from the touch sensing operation pattern database as compared with FIG.

Touch sensing pattern codes 0 to 10
In this embodiment, as the selection of the touch sensing operation pattern corresponding to ..., the touch sensing pattern table addresses at0 to at10 ... and the number nt of the sensing pattern table data (see Fig. 4) are determined. Here, the touch sensing pattern table address at
0 to at10... Are the touch sensing operation patterns corresponding to the touch sensing pattern codes 0 to 10 (for example, the touch sensing operation pattern 18 in FIG. 4).
Indicates a position (address) stored as a table in the teaching database 16. The number nt of the touch sensing pattern table data is the number of touch sensing operation patterns in the table. In the touch sensing operation pattern 18 in FIG. 4, the number of sensing pattern table data is four.

Next, in step S12 of FIG. 2, a sensing operation is automatically generated. That is, based on the touch sensing pattern table addresses at0 to at10... And the number of touch sensing pattern table data nt.
The teaching data 12 of the touch sensing operation is generated.

In the examples of FIGS. 3 and 4, the touch sensing operation pattern 18 can be registered for each touch sensing pattern code 0 to 10... is there.

The touch sensing operation pattern 18 shown in FIG. 4 shows the simplest touch sensing operation of horizontal welding of two fillets. Here, the touch sensing operation pattern 18 will be described.
2, L3), forward position (L
1, 0, L3), a retreat position (L1, L2, L3), a descending position (L1, L2, 0) for touch sensing, and an ascending position (L1, L2, L3). The reference position, the retreat position, and the ascending position are the same. As shown in FIGS. 5A and 5B, each position is represented by an XYZ orthogonal coordinate system having a point A on a welding line as an origin.

Specifically, as shown in FIG. 5A, first, from the reference position (L1, L2, L3), Y =
0, the tip of the welding torch 19 is moved to perform touch sensing on the surface of the standing rib member 20, and from there, return to the reference position, and move downward to Z = 0 to perform touch sensing on the surface of the member 21; A series of touch sensing operations of returning to the reference position are defined.

The touch sensing operation pattern 18 shown in FIG.
In this case, the posture of the tip of the welding torch at this time and the sensing speed can be specified. The attitude of the welding torch tip is two angles α in the polar coordinate system,
It is represented by β.

The database of the touch sensing operation pattern 18 and the like and the touch sensing pattern code 0
The database 17 of 10 to 10 allows arbitrary definition of various member shapes and touch sensing operations of each welding robot maker.

As described above, the teaching data 12 incorporating the touch sensing can be automatically generated for each welding line of the member (or for each unit of the member).

Therefore, if the welding line n is processed to the end as in step S13 of FIG. 2, finally, the teaching data 12 of the welding robot can be automatically created offline for all designated members.

Then, the teaching data 12 is
The data is transferred to and registered in the memory of the robot controller via a communication line, or recorded on an external recording medium such as a floppy disk, taken from the office to the robot controller, and registered in the memory. Welding is performed using the registered teaching data 12.

At this time, in the touch sensing operation included in the teaching data 12, the welding robot actually moves and touch-senses the member position, and the position of the welding robot is corrected by assuming the member position, and the welding line is shifted. And perform welding.

Alternatively, a three-dimensional operation simulation 14 of the welding robot is performed on the graphic display based on the teaching data 12, and the appropriateness of the touch sensing operation and the welding operation is confirmed by referring to the interference check list 15 (see FIG. 1). .

The three-dimensional off-line automatic teaching system 13 in FIG. 1 comprises a computer (CPU: central processing unit) and a computer program (software) necessary for realizing the present invention. In order for the computer to execute the processing shown in FIG. 2, the data is recorded in an appropriate recording medium in the form of data readable by the computer.

Although one embodiment of the present invention has been described above, the present invention is not limited to the above embodiment, and various modifications can be made within the technical idea of the present invention. Also belong to the technical scope of the present invention.

For example, the present invention is not limited to touch sensing of a wire contact type or the like, but may be a non-contact sensing using a laser beam, a camera, or an electric sensor such as an eddy current sensor or a capacitance sensor. In the same way, the off-line automatic teaching and the sensing operation simulation can be performed for the recognition of the start and end of the member and the recognition of the groove shape.

For touch sensing or non-contact sensing of not only welding robots but also various robots such as painting robots, offline automatic teaching and sensing operation simulation can be performed by the same method.

[0061]

According to the sensing pattern offline automatic teaching method of the present invention, it is possible to arbitrarily define different touch sensing or non-contact type sensing for each of various member shapes and robots. It can be created with high accuracy and efficiency. In addition, a simulation of sensing operation can be added, and interference check at the time of sensing can be performed.

[Brief description of the drawings]

FIG. 1 is a diagram illustrating an entire system configuration according to an embodiment of the present invention.

FIG. 2 is a view for explaining an outline of a processing flow in one embodiment of the present invention.

FIG. 3 is a diagram illustrating an example of a touch sensing pattern code in a touch sensing pattern code database according to an embodiment of the present invention.

FIG. 4 is a diagram illustrating an example of a touch sensing operation pattern in a touch sensing operation pattern database according to an embodiment of the present invention.

FIG. 5 is a diagram showing an example of a touch sensing operation in one embodiment of the present invention.

[Explanation of symbols]

0-10 Touch sensing pattern code 11 Design CAD data 12 Teaching data 13 3D offline automatic teaching system 14 3D operation simulation 15 Interference check list 16 Teaching database 17 Touch sensing pattern code database 18 Touch sensing operation pattern 19 Welding torch 20 Rib member 21 Member A Point on welding line at0 to at10 Touch sensing pattern table address nt Touch sensing pattern table data number

──────────────────────────────────────────────────の Continued on the front page (51) Int.Cl. ⁶ Identification code FI B23K 9/127 508 B23K 9/127 508A 509 509B B25J 9/22 B25J 9/22 A 19/02 19/02 G05B 19/42 G05B 23/02 301L 23/02 301 19/42 J

Claims (5)

[Claims] 1. A sensing operation pattern database that defines a plurality of types of sensing operation patterns different for each member and a shape of a member, and a plurality of types of sensing pattern codes corresponding to each sensing operation pattern in the database. And a sensing pattern code database defined by a key such as a constraint condition of a robot, a specific sensing pattern code is selected from the sensing pattern code database by the key, and a sensing operation pattern corresponding to the specific sensing pattern code is selected. The teaching pattern is selected from the sensing operation pattern database, and the teaching data is automatically created offline according to the selected sensing operation pattern. Teaching method. 2. The method of claim 1, wherein the sensing is touch sensing. 3. The sensor according to claim 1, wherein the sensing is a non-contact sensing using a sensor using a laser beam, a sensor using a camera, or a sensor using electricity such as an eddy current sensor or a capacitance sensor. 2. The method for automatically performing off-line teaching of a sensing pattern according to claim 1. 4. The offline automatic teaching method of a sensing pattern according to claim 1, wherein the sensing is performed by a welding robot to recognize the start and end of a member, to recognize a groove shape, and the like. 5. A sensing operation simulation method, wherein a sensing operation of a robot is simulated on a graphic display by using teaching data created by the offline automatic teaching method for sensing patterns according to claim 1. .