JPH0366577A - Preparation of teaching data for plural robots - Google Patents

Abstract

PURPOSE:To quickly obtain a teaching data having a short working time and well balanced, in the case of the selection of a robot for the teaching point of the work area inside where the motion zones of two robots being superposed, by deciding the work allocation of a teaching point so as to select the robot whose working time accumulated at that time is shortest. CONSTITUTION:The selection (P1-2) of a robot at the teaching point of the part where the work area of a robot is not superposed is performed by selecting a corresponding robot according to the order set in advance, for example, or selecting a corresponding robot with a specific rule. Work is executed (P1-3) by simulation by the robot thereof and the work time thereof is accumulated on each robot (P2). The selection (P1-1) of the robot at the teaching point of the part where the work area of the robot is superposed is performed with the work time of each robot already found by the calculation process of the work time executed repeatedly as the reference.

Description

[Detailed description of the invention] [Industrial field of application] One aspect of the present invention: For example, when multiple robots are operated at the same time to perform work, the area where the operating areas of the robots overlap may be reduced. This invention relates to a method for creating teaching data for multiple robots that determines the distribution of work.

[Prior Art] Conventional Work Efficiency Method Welding robots and painting robots are used on automobile production lines and the like from the viewpoint of productivity or safety and health. These robots (friends)
Since it is possible to repeat the same work accurately, once appropriate teaching data (operation program) is taught, it is possible to continuously obtain products with high and consistent quality (Japanese Unexamined Patent Publication No. 180257/1983). .

Furthermore, from the perspective of improving work efficiency and productivity, multiple robots may be operated simultaneously. In this case, it is necessary to make each robot share the work. In order to save labor in this case, a method has been proposed (Japanese Unexamined Patent Publication No. 63-295192) in which the division of labor is automatically done. This method is carried out as follows. First, work priorities are set for each of the two robots, and the robots are made to perform the work through simulation.

If tasks overlap, the robots with the highest priority are assigned to work first. The simulation is repeated for this task by changing the combination of priorities, and the priority task with the shortest working time is selected from among the obtained working times.

[The problem that the invention aims to solve] However, this method requires a relatively long time because it measures the work time by repeating the simulation work many times while switching priorities between robots. For example, even if the number of robots is 2 (L), at least 2 (2!) simulation processes are required, for 3 robots it is 6 (3 times), and for 4 robots it is 24 (4
j) Simulation processing is required several times. Moreover, as the number of robots increases, the number of simulation tasks increases rapidly.

Furthermore, although the selection criterion after the simulation is "the shortest work time," the parameters that can be changed in the simulation are the priorities of the robot. Furthermore, during each simulation task, each task is performed with a fixed priority order. Therefore, since detailed consideration has not been made for each task, it is rare that a well-balanced allocation of work time to each robot is achieved, and even if all combinations of priorities are executed, the obtained work time data may There is no guarantee that there is a true shortest working time or a shortest working time.

Therefore, it is not efficient to create a motion program as teaching data (:, it is almost impossible to obtain a motion program that minimizes the work time by distributing the work time of each robot in a well-balanced manner.) However, the repeated distribution work through trial and error consumes a huge amount of time and effort, and it is not possible to obtain an efficient operating program.

Therefore, in the conventional method, the quality of the motion program was poor and the creation efficiency was low, and the present invention was aimed at solving the above problems in the operation of multiple robots. [Means for Solving the Problems] That is, the gist of the present invention is to adopt a configuration in which each work area is arranged so that at least a part of it overlaps, as illustrated in the basic configuration diagram in FIG. This method creates teaching data for multiple robots to work at the same time by associating multiple robots with teaching points set in a work area. Calculate the working time for each robot by (P2) and calculate the robot's working time for the teaching point in the work area where the motion ranges of at least two robots overlap. When choosing
The method of creating teaching data for a plurality of robots is characterized in that the robot with the shortest working time is selected at the time of selection (PI-1).

[Operation] Selection of robots at teaching points in areas where robot work areas do not overlap (PI-2) I For example, select the applicable robot according to a preset order, or select the applicable robot according to a predetermined rule. done by doing. Then, the robot is made to perform the work in simulation (PI-3), and the work time is accumulated for each robot (P2).

Selection of robots at teaching points in areas where robot work areas overlap (Pl-1) Ru. That is,
This is done by selecting the robot with the shortest working time.

Therefore, the teaching points and the robots that perform the work are associated in such a way that the accumulated work time is always uniform, and in this way the teaching data (movement program) that is short and has a good balance of work time can be quickly created. can get.

[Example] Next 1: Detailed explanation of the present invention The present invention is not limited to these, but includes various embodiments without departing from the gist thereof.

FIG. 2 is a system configuration diagram showing an example of a simulation device that implements the method of the present invention.

Simulation device 1 (top) It is mainly composed of a general Neumann type digital computer 3. This computer 3 has a CPU and a ROM.

It is composed of RAM, Ilo, bus, etc., but since it is a general configuration, a detailed explanation of the inside will be omitted. This computer 3 (friend) has CR as an output device.
T5. The printer 7 uses a floppy disk device 9 as an auxiliary storage device. A keyboard 13 is connected to the hard disk device 11 as an input device.

Procedure of the embodiment method of the present invention 1 Company R of the present computer 3
It is read as a program from the hard disk device 11 during AM. In accordance with this program, the computer 3 reads the position and orientation data of the teaching points stored in the floppy disk device 9, and further reads the operational performance data and shape data of each robot stored in the hard disk device 11. , the robot to perform the work at the teaching point is selected based on the shortest cumulative work time. The motion of the robot's arms, etc. is then simulated using calculations, and the robot's working time is accumulated. The position/orientation data of the teaching point mainly consists of data representing its three-dimensional coordinates, the position of the welding torch at the tip of the arm (positional data), and the direction of the welding torch (orientation data), and the following data is also added. ing.

For example, the IL teaching point is given along the workpiece W (in this case, the car body) at a distance of 1 m, as shown by the black circle in Fig. 3 <a>. Also, the data structure for each teaching point (
As shown in FIG. 3(b), it is composed of position/orientation data d1 on three-dimensional coordinates, its taught false name (taught point NQ) d2, +Dd3 of reachable robots, and occupancy flag d4. IDd3 represents a code for robot identification,
The occupancy flag d4 represents a flag to which the ID of the robot selected in the work distribution process (FIG. 6), which will be described later, is assigned.

i1Dd3 and occupancy flag d4F upper This data is determined in the work allocation process described later, so it is not initially set.

In the simulation device 1, a process is performed in which the corresponding robot sequentially traces the teaching point position of one company while performing welding or spraying work. Then, each time the work at each teaching point is completed, the working time of the teaching point is accumulated in the working time of the corresponding robot.This accumulation is repeated until all teaching points are completed, and the total working time of each robot is obtained. During calculation processing, three-dimensional images of the workpiece and robot can be sequentially displayed on the CRTS according to instructions from the keyboard 13 according to simulation calculations.

C, which performs simulations based on the operational performance data and shape data of multiple robots, and the teaching point data to be passed (position data, nozzle direction data, etc.)
AD program (above) A commercially available program can be used, for example, IBM's product name rCAT I AJ.
, Product name rCIM-8TAT manufactured by Mitsui Engineering & Shipbuilding evI
One example is 10NJ.

Next, the welding robot to which the operation program (teaching data) obtained by the above-mentioned simulation device is applied,
The arrangement within an automobile body welding line is shown in Figure 4.

Two welding robots 17 and 19 are placed on the engine room 15a side of the automobile body 15 that has been moved on the line and brought into the booth, and two left and right welding robots 17 and 19 are placed on the rear luggage room 15b side. Robots 21.23 are placed. Each robot 17 to 231 is equipped with a robot control device 31, 33° 35.37, and moves its arm according to the operation program stored in the floppy disk loaded in each disk device 31a, 33a, 35a, 37a. It operates the welding torch and stores data taught by the operator via the teaching box on a floppy disk.

Motion range 1 shown by the dotted line of the front two robots 17.19
78.19Sl& Each covers the engine room 15a side. Rear robot 21°23 operating range 2
The same applies to 1S and 23S. Their operating ranges 173-23S partially overlap.

The welding robots 17 to 23 are well-known articulated robots, consisting of a skeleton part, a first arm, a second arm, and a welding torch, and each of these skeletons can perform movements such as turning by a predetermined drive part. It is possible.

The robots 17 to 23 have operating performance as shown in Table 1. Phrases: Robots with various operating performance can be individually selected and used depending on the purpose.

The data in Table 1 in Table 1 and Table 2 are stored in the hard disk drive 1 of the simulation device 1 as part of the operating performance data together with the shape data.

The teaching work for the body 15 is performed by the robot control devices 31 to 31.
37, and the teaching boxes 31b to 37b.

In this embodiment, the operator selects in advance a suitable robot 17-23 for the welding part of the body 15 and performs a teaching operation, and the data is stored in a floppy disk in the corresponding disk device 31a-37a. If it is stored on a floppy disk, it is converted to one common coordinate and all teaching points are created as common position/orientation data.

As another method of creating position/orientation data, if shape data of the body 15 already exists, teaching points may be determined by simulation or simple calculation based on that data, and position/orientation data may be created.

By creating the position/orientation data as described above, it becomes possible for the simulation device 1 to execute a simulation.

Below, regarding the work allocation process by simulation,
This will be explained based on the flowchart shown in FIG. Processing of this flowchart is executed by reading a program stored in the hard disk device 11 into the RAM of the computer 3.

When the simulation device 1 is started up, a program for work allocation processing is read from the hard disk device 11 and processing according to the program is then started.

First, the necessary data is read into the work area in the RAM (step 90), and the necessary data and the shape and operating performance data of various robots already stored in the hard disk drive 11 are used for this simulation. These are the data on the robot model to which the data is to be assigned, the shape data of the automobile body 15, and the position/orientation data for each teaching point shown in FIG. 3(b), which have already been obtained in the teaching work.

If there is position/orientation data as shown in Figure 3(b),
C The robot obtained by teaching the data and the robots 17 to 23 to be simulated this time may be different. In this respect, data representing position/orientation has versatility. That is, even if the model of the robot is inconvenient as a result of the simulation (fS), it is possible to immediately change to another model and perform the simulation (1i), which is also useful in determining the model.

If the data format on the robot control devices 31 to 37 side does not match the data format on the simulation device 1 side, 1iWhen reading the data into the RAM, a format conversion process is performed and the data is read.

Next, for each read teaching point, each robot 17 to 23 determines whether or not it is an reachable teaching point where welding work can be performed.
Each robot is determined by simulation or simply by whether it is within a known reachable area for each robot (step 100). In this case, simply robot 17~
Not only the movable coordinate position of the welding torch 23, but also the posture of the arm and welding torch, interference between the robot 17 to 23 and the body 15, or the position of the arm and its tip so that the welding work can actually be performed at the teaching point. It is determined whether or not it is reachable, taking into account interference with the welding torch.

Specifically, the robots 17 to 23 perform the inverse coordinate transformation (conversion from target position and posture data to the angle of each axis of the robot) and determine whether each axis value is operable or not. This is done by determining the positional relationships of individual objects such as the arms of the robot, tools such as welding torches, tool brackets, and workpieces (body 15), and checking for interference.

As a result, the IDs of all the robots that reach the robot are entered in the data array of reachable robot IDs in the data set for each teaching point shown in FIG. 3(b).

For example, assume that the teaching points are arranged as shown by black circles in FIG. IDs (A, B) of robots 17 to 23 that can reach the data of all teaching points.

If you fill in C and D), it will look like the ID written along each black circle. That is, only the ID of I''A4 is written at the teaching point that can only be reached by the robot 17.2
The teaching point that can only be reached by robot 17.19 is r.
The IDs of AJ and rBJ are entered. In this way,
Depending on the robots 17 to 23 that can be reached, one or more IDs are entered in the position/orientation data of the teaching point.

The contents of Fig. 7 may be displayed directly on the CRT 5 as a plan view.Furthermore, in the display, the teaching points and D are color-coded depending on the type of robots 17 to 23 that can be reached, so that the operator can identify them. You can make it easier.

Next, the teaching point data is classified (step 200), for example, by the ID written in the position/orientation data. The teaching points are classified into teaching points whose ID is only "A", teaching points whose ID is only "A" and rBJ, and so on. In the example of Figure 7 (9 Classification IE, [A], [B] [C]
, [D], [AB], [AC],
[AD] [BD], [CD], [ABC]
, [ACD]. Each region (teaching point set) representing this type is hereinafter referred to as a "module", and is sequentially distinguished by symbols M1 to Mll.

Next, a work route is set for each module M1 to M]1 (step 300). The working path is to order the teaching points. For example, starting from the end of the teaching point array on the coordinates, the path is set by tracing adjacent teaching points.

You can form a route by giving priority to the branching part (above), and follow the branching part (Yo) after completing the straight-ahead part, or you can give priority to the branching part that is not going straight, and When you reach the end of the branch, you may return to the root of the branch and follow the straight part.Furthermore, instead of having Computer 3 process everything, when you reach a branch, you can use one friend.Computer 3 The computer 3 may request the operator to select a branch, or the computer 3 may request the operator to input the entire route. The route is set.In addition, in Fig. 8, the teaching point circled in each area is the starting point or the ending point.
Only the teaching point is given a code. It is assumed that codes are continuously assigned to the teaching points between the starting point and the ending point. Further, the arrows simply indicate the continuous state of the route, and not only the forward direction of the arrow but also the reverse direction can be selected as the work route.

Next, the operator is requested to create a preliminary operation program for each of the modules M1 to M11 that operates along the route for each robot 17 to 23 (step 400).

Based on this request, the operator creates one preliminary operation program for each of the corresponding robots 17-23 for the modules M1-M4.
Regarding Mll (Ambition) Multiple preliminary movement programs are created for each reachable robot.

For example, in module M10, which is shown enlarged in Figure 9, it is a common work area for robots 17, 19, and 23, so three preliminary motion programs are created for each of the robots 17, 19, and 23. .

A preliminary motion program for the robot 17 starting from li TplO-7 and ending with T PIG-1 is created as follows. First, move in a straight line to the teaching point T I) 10-7 at speed Val and acceleration Aal.
■Transfer in a straight line to teaching point T 910-6 at speed Va2 and acceleration As2 ■Welding operation ■Speed V to teaching point TplO-5
A3, moving in a straight line with acceleration As2 ■Shout...'' A motion program is created in which motions for each teaching point are combined. It may also be written using a general robot language. However, since the preliminary operation program (see below) is selected for each teaching point in the work distribution process (step 600) to be described later, the program structure should be such that it can be deleted, added, or changed in units of teaching points.

A preliminary motion program is also created for the robot 19.23 from the teaching point Toto-7 to the teaching point TplO-1. Low Depending on the arrangement of the robot, the work order of the modules, etc. (on the contrary, a preliminary operation program from teaching point T plO-1 to teaching point T plO-7 may be created. For example, module Ml1, etc. .23 and robot 21 have opposite operating directions selected.

Also, robot 17. 19. Due to the arrangement of 23, there is an obstacle F on the way and the teaching point TplO-7-Tpl
If it is not possible to move between O-6 and 1, an escape point Hp is provided as shown in Figure 9, and the teaching point TplO-7-7°1 is set.
A preliminary operation program may be created so that when moving between 0 and 6, the escape point Hp is passed through.

In this way, a preliminary operation program is created for each reachable robot 17-23 for all modules M1-Mll. The preliminary operation program can be created not only by the operator himself, but also by the computer 3 automatically.

IL: Computer 3 is already robot 17-23
Since performance data has been obtained, the motions of the robots 17 to 23 may be simulated according to the performance data, and if there are no problems such as interference, the motions may be automatically registered as a preliminary motion program.

Additionally, the operator may be able to observe this simulation on the CRT 5, and if there is a defective operation, the operator may instruct the operator to change the operation conditions and re-simulate instead of adopting it as a preliminary operation program. .

Next, for each robot 17-23, the work order of the work target modules M1-Mll is determined, and in that order, the preliminary operation program for each module is transmitted to the robots 17-23.
3 to create a series of preliminary operation programs (step 5oo).

Each of these robots 17 to 23 (integrated preliminary movement program structure S^, s s, s c, s
o is shown in FIGS. 10(a) to (d).

That is, for robot 17 (upper 6 modules/L
, Ml, M5. M6. M7. Mho and M17 are executable. Thus, for example rMl-MS-.

It is decided to execute in the order of M 10 - M 7 - M 11 - M 6 J, and robot 1 is created for each module.
The six preliminary operation programs of No. 7 are integrated into a series of preliminary operation program structures S^ as shown in FIG. 10(a). Also robot 191; rM 2-
M S-M 10-M 8J, rM3-M6-Ml 1-M9J for Rohot 21, rM4-M8-M 10-M7-M 11-M9J for Robot 23
The execution order of the modules is determined, and the first
Figure 0 (b), (C).

A series of preliminary operation program structures S as shown in (d)
Unify as B, SC, and SD.

Note that 1IkL (hereinafter referred to as program floor) in each structure S^~SD represents a consecutive number assigned to a unit operation program for each teaching point, and the order is as follows for each structure S A -
It represents the execution order in SD.

The execution order of the modules 1t may be determined by the operator while looking at the display on the CRT 5, but the computer 3 may sequentially select the end point and start point, which are the shortest distances in the robot operation, from different modules.
The execution order may be determined by determining the connection state of the modules. However, only the modules M1, M2. M3. Regarding M4 (L Since only that robot can work, each structure S A, S B, S C, S D
Select as your first task.

The structures S A to S D are composed of modules M1 to M]1 arranged as described above, and unit operation programs and occupancy flags for each teaching point therein. This arrangement order corresponds to the execution order.

However, as mentioned above, the occupancy flag (at the time the preliminary operation program is created) is undecided.

Further, the occupancy flag is attached to the teaching point, and a common ID is given to the same teaching point even if the structures are different. Therefore, for example (e) rAJ is set at the teaching point TplO-7 (program NCL37) of the module M10 in the structure SA of FIG. 10(a) set for the robot 17.
is set, "A" is also set at the same time at the teaching point TplO-7 (program stack 21) of the module MIO of the structure 'SR in FIG. 10(b), which is set for the robot 19. At the same time, rA,1 is also set at the teaching point TplO-7 (program Nc16) of the module M10 of the structure SD in FIG. 10(d).

Next, the process (step 600) of allocating work to each robot 17 to 23 and determining the occupancy flag will be described.

The details of this work distribution process are shown in the flowchart of FIG. 6. First, four end flags Fel, Fe2, . F e3
．． F e4 to each structure S A, S B, S C
, S D (step 605), that is, FIGS. 10(a), (b), (C), (
d) Program teaching point 1 T pi-25, T p
2-1. T p3-1. Set in the T p4-1 section. Next, it is determined whether or not there is a teaching point for which the occupancy flag has not been determined by searching all four structures 5A-8D.

If there is even one teaching point for which the occupancy flag has not been determined, an affirmative determination is made.

Initially, since the occupancy flags have not been determined for all teaching points, an affirmative determination is made and the robot with the shortest accumulated work time is then selected. The cumulative work time is set in the memory in the RAM for each robot 17-23. Initially, the cumulative work time of all robots 17 to 23 is rQJ, so
A robot to perform the following processing is selected in a preset priority order. In this embodiment, robot 17 has the highest priority, and the following priority orders are set in the order of robots 19, 21, and 23.

Therefore, the robot 17 is selected and the following processing (steps 620 to 675) is performed.

First, the content of the occupancy flag of the teaching point of the structure SA pointed to by the end flag Fel is determined (step 620
), even if its occupancy flag is "0", it is set to IDrAJ of the robot 17 (step 625).

Next, the cumulative work time of the robot 17 is updated (step 630), that is, the work time from the current posture/position of the robot 17 to the point where it is moved to the teaching point and the work there is completed is updated in Simulation 1. Requested tag The time will be added to the cumulative work time accumulated up to the previous time.

Next, the end flag Fel is updated by one step (
Step 635). For example, if the end flag Fel points to the program F4CL1 (teaching point T pi-25) of the structure SA, the end flag Fel points to the program F4CL1 (teaching point T pi-25).
pl-24).

In this way, the process returns to step 610 again.

Next, after an affirmative determination is made in step 610, step 61
In Step 5, one robot with a high priority is selected from the robots 19, 21, and 23 whose current cumulative work time is the shortest (in this case, rO''), and similarly steps 620, 625, and 6 are performed.
In the process of 30,635, the program positive 2l (Ta2-
rBJ is set in the occupancy flag of 1). The accumulated work time is added. The end flag Fe2 is updated and processing is performed to point to program N112 (Ta2-2).

Thereafter, no processing is performed for the robots 21 and 23 as well.The above processing is repeated for the robots 17 to 23 selected according to the shortest cumulative work time and priority, so that each structure S This is a representative example of how the occupancy flags of teaching points existing in the motion range of a plurality of robots are set in the work distribution process (step 600) after the occupancy flags of A to S D are sequentially determined. Let's explain with an example.

[1] First, regarding robot 17, the occupancy flag is the 10th
As shown in Figure (a), the end flag Fel points to the teaching point T p5-11 of the program positive 26, and the occupancy flag for the robot 17 is shown in Figure 10 (
It is assumed that the end flag Fe2 points to the teaching point Tp5-8 of the program 13.

In this state, consider a case where an affirmative determination is made in step 610 and the robot 17 is selected in the process of step 615. Then, in the process of step 620, the end flag F
rBJ has already been set in the occupancy flag of the teaching point T p5-11 pointed to by el. Therefore, a negative determination is made, and then the teaching point data pointed to by the end flag Fel is not at the beginning of the module, and the occupancy flag of the previous step (in this case, Tpt-t of program 25) is the same as the code rAJ of the robot 17. It is determined whether or not (step 640). Here, since the end flag Fel is the beginning of the module M5, a negative determination is made, and then the end flag Fel is advanced by one (step 645), which points to the teaching point T p5-10 of the program N1127. Nagi If the previous end flag "el is the last program positive 55, go up by 1 at step 645. The end flag Fel indicates the end"
-2'' value is set.

Next, it is determined whether or not the end flag Fe1O value is "-2" (step 650). 27, a negative determination is made, and it is determined whether the occupancy flag of the teaching point Tp5-10 in the program Nn27) at the end flag Fe1 position is not "0" (step 655).
). Since "B" is further set here, a negative determination is made and the end flag Fet is updated by one again (
Step 645), that is, the program 28 is pointed to.

Similarly, r13J is added to the occupancy flag of program N1128.
is set, the end flag Fel is further updated by one (step 645), and points to program Nn29. Therefore, a negative determination is made in step 650.Furthermore, since the one company occupancy flag in this program 29 is rOJ, an affirmative determination is made in step 655, and the occupancy flag is set to 1DrAJ of the robot 17 (step 625).

Next, the robot 17 is simulated at the teaching point Tp.
Move and work from l-l to teaching point T p5-8,
The time is measured and added to the cumulative work time of the robot 17 to accumulate it (step 630).
el (step 635), i.e. program N
n30, and the process returns to step 610 again.

[2] The processing when an affirmative determination is made in step 640 will be explained using the processing of module M10 as an example.

Consider a case where the work distribution process has progressed to a certain extent and the state shown in FIG. 11 is reached, and it is determined in the process of step 615 that the robot 17 has the shortest accumulated work time. still,
FIG. 11 (1), (2), and (3) 1 and 1 respectively show only the module MIO portions in the structures SA, SB, and SD.

At this time, in step 620, the end flag Fel of the robot 17 is set to the program memory 38 shown in FIG. 11(1).
, and its occupancy flag is "rB", so a negative determination is made.

Next, the program floor 38 (TplO-6) pointed to by the end flag Fel is the beginning of the module MIO (7
, 10-7), and program floor 3 of the previous step.
7 (teaching point T plO-7) occupancy flag is robot 1
It is determined whether or not it is 1DrAJ of 7 (step 6).
40). Here, since the condition is satisfied, an affirmative determination is made.

Next, the still occupied flag in that module M10 is “O”.
” (step 660).

Since the occupancy flag of the program N1141 in the module M10 is "0", search for a teaching point assigned an ID in the same module MIO, after the program floor pointed to by the end flag Fel in the 0th order where an affirmative determination is made. . The teaching points are program Nct38-40 (T p
lo-6, T plO-5, T plO-4).

And the ID of the occupancy flag of the program magnetic fields 38 to 40
For the robot 19.23 corresponding to the content, the teaching point T pi (1-6, TpiO-5,
Subtract the working time of TplO-4 (step 66
5). This is to move the ID and return the accumulated work time of the robot 19.23 corresponding to D to the state immediately before that work.

Next, set the D contents of the occupancy flags of programs 38 to 40 to 1.
Shift backward by one (step 670), and the teaching point T plO-5, T pl to which the ID has moved during the accumulated work time of the robot 19゜23 corresponding to the next shifted ID.
The working time of O-4, T plO-3 is calculated by simulation and added (step 675), thus correcting the cumulative working time of robot 19.23 according to the ID assigned to the occupancy flag.

Next, the occupancy flag of the program floor 38 pointed to by the end flag Fel is set in l Dr AJ of the robot 17 (step 625).

When the state of the occupancy flag of the structure S^ of the robot 17 at this point is shown, it can be seen that DrAJ has been set continuously as shown in FIG.

Next, the teaching point T is set at the cumulative time of the white box 11F0 work.
The working time of plO-6 is added (step 630);
The end flag Fel is updated by one step (step 635). That is, the end flag Fel is the program N11.
69 (TplO-5).

[3] Module M as shown in Fig. 13 (1) to (3)
Assume that the robot 21 is selected as the robot with the shortest cumulative work time when the occupancy flag of ll is set and the end flag Fe3 of the structure SC is set to program N1141.

At this time, the occupancy flag of the program N116 pointed to by the end flag Fe3 is set to rDJ, and a negative determination is made in step 620, and the program target 16 (Tpll-
3) is not the beginning of the module M11, and it is determined whether the occupancy flag of the program m 15 (T pH-2) in the previous step is the same as IDrCJ of the robot 21 (step 640). The program Nn16 (Tpll-3) pointed to by the end flag Fe3 is the module M7.
Since the IO at the previous program floor 15, not at the beginning of program level 1, is rCJ, an affirmative determination is made. In the next step 660, a negative determination is made because IDs are set in all the occupancy flags in the module M11.

Next, the end flag Fe3 is updated by one step (step 645). That is, the end flag Fe3 points to the program ml7 (Tp9-11). Next, it is determined whether the value of the end flag Fe3 is "-2" (step 650), that is, it is determined whether the arrangement of the structures SC of the robot 21 has been completed. Since the array is not completed, a negative determination is made and it is determined whether the occupancy flag is "O" (step 655). Since the occupancy flag of the program pile 17 is rQJ, an affirmative determination is made and the occupancy flag is set to "C" (Step 625), and the movement from the teaching point TpH-2 processed immediately before to the teaching point Tp9-11 is performed. The work time is determined by simulation and added to the cumulative work time of the robot 21 (step 630). In this way, the end flag F
After updating e3 (step 635), that is, pointing to program Net18 of structure SC, the process returns to step 610.

The above-mentioned process is repeated, and finally the image shown in Fig. 14 (
Structures SA, SB, SC, and SD as shown in a), (b), (c), and (d) are created.

As explained above, in the process of step 600, the work at the teaching points is distributed so that the cumulative work time is evenly distributed.
5. Through the series of processes 670 and 675, the teaching points in the common work area are grouped for each robot, making it possible to allocate the teaching points for efficient work.

As shown in step 2, all teaching points with undetermined IDs are structures S A
- cease to exist in SD (ヱ Step 61
14(a) created next after a negative determination was made at 0.
A teaching operation program for each robot is created from each of the structures S A to S D shown in (d), and its registration and modification processing is further executed (step 700). That is, for the robot 17, only the data of the teaching points of the structure S^ whose train occupancy flag is "A" are sequentially retrieved and integrated. be done.

The data is stored on this floppy after data conversion is performed as necessary.

In addition, if a simulation is executed using the teaching motion program created in this way and an inconvenient motion occurs, for example, if interference with another robot or interference with the body 15 that was unknown until now is discovered, the teaching motion program will be used. A process is performed to modify the operating program. Other robots 19-2
Similarly for 3, structures S B, S C, S
A teaching operation program is created, registered, and modified from D.

In this way, the simulation (teaching motion program) is completed.The teaching motion program is then read into the control devices 31-37 of the robots 17-23 and used for actual work or trials. As a result of this work, if there is any inconvenience, the control device 31~
37, or the teaching operation program may be returned to the simulation device 1 again and the teaching operation program may be corrected in the same manner as in step 700 while checking the operation through simulation. good.

The above was about work distribution among four robots, but the same distribution is possible with two to three robots, or five or more robots.

In the method of creating teaching data for multiple robots according to the present invention, a robot is selected for each teaching point and the working time is accumulated for each robot, so that the operating ranges of at least two robots overlap. When selecting a robot for a teaching point in the work area, the robot with the shortest accumulated working time at that point is selected to determine the work distribution for the teaching point, so each robot's work is always The teaching point distribution is set so that the time is uniform.In this way, the working time is short and well-balanced teaching data (operation program) can be quickly obtained.

[Brief explanation of drawings]

FIG. 1 shows an example of the basic configuration of the present invention. FIG. 2 shows the system configuration of an embodiment. FIG. 3(a) shows an explanation of the distribution of teaching points along the workpiece (automobile body). ) is the data structure explanation of each teaching point. Figure 4 is the arrangement of the welding robot. Figure 5 is a flowchart showing the work allocation process executed by the simulation device. Figure 6 is the flowchart showing the work allocation process. Flowchart, Figure 7 is an explanation of the distribution of teaching points with respect to the robot's placement, Figure 8 is an explanation of the path of the teaching points, Figure 9 is an enlarged explanation of one of the modules, and Figures 10 (a) to ( d) is a structural explanation of the preliminary motion program structure set for each robot.
) to (3) are partial explanations of the preliminary operation program structure. FIGS. 14A to 14D are explanatory views of the completed state of the preliminary operation program structure. 1... Simulation device 17, 19, 21. 23...Robot 15...
Body (work) 31, 33, 35. 37... Robot control device M1
~Mll...Module 17S-23S...Movement range Tp1-1 of each robot
~T pH-3...Teaching point S A - S D

Claims (1)

[Claims] Teaching data for multiple robots to work simultaneously is provided by associating a plurality of robots arranged so that at least a portion of each work area overlaps with a teaching point set within the work area. This method calculates the working time for each robot by selecting a robot for each teaching point and having it work in a simulation, and also calculates the working time of each robot and makes sure that the operating ranges of at least two robots overlap. A method for creating teaching data for a plurality of robots, characterized in that when selecting a robot for a teaching point in a working area, the robot with the shortest working time at the time of selection is selected.