JPH0699381A - Control method for robot provided with visual sensor - Google Patents

Abstract

PURPOSE:To improve work efficiency of a robot provided with a visual sensor, relating to a control method for the robot provided with the visual sensor. CONSTITUTION:By a visual sensor-provided robot 5 provided with a work tool 2, and a visual sensor 1, which can be rotated with an axis of the work tool 2 serving as the center in a robot head part, the visual sensor 1 passes to go ahead of the work tool 2 above a work position on a workpiece 3. A peripheral part of the work position is photographed without stopping the robot during moving, to obtain the actual work position by detecting a position of the recognized object from a picture data, and the work tool 2 is moved to the actual work position so as to perform work.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for controlling a robot with a visual sensor, such as a screw fastening robot with a visual sensor. Hereinafter, the robot will be described as a screw tightening robot.

[0002]

2. Description of the Related Art A conventional screw tightening robot moves a work tool such as an electric screwdriver provided in a head portion to a work position such as a screw pilot hole on a work to teach work position data of the work tool and operate the work tool. At this time, the work tool is positioned by the positioning control device according to the teaching work position data, and the screw tightening work is executed. However, in this method, for example, in the case of a plastic molded product or a work of a sheet metal processed product having a large outer dimension, in which the molding dimension is easily influenced by external factors such as molding conditions and temperature, the work tool is accurately positioned at the teaching work position. However, since the actual position of the screw pilot hole on the workpiece may be displaced from the teaching work position, reliable screw tightening work cannot be performed, making automation of the screw tightening work difficult.

Therefore, in order to solve the above problems,
A visual sensor is attached to the head part, the visual sensor is moved to the screw hole position on the work to teach the position, and during operation, the visual sensor is temporarily positioned and stopped at the taught position, and then recognized by the visual sensor. The actual work position is obtained by detecting the position of the center of gravity of the screw hole that is the object, and the corrected movement amount to which the work tool should move is calculated from the detected work position data and the mounting position relationship data of the work tool and the visual sensor. Then, the work tool is moved according to the corrected movement amount, and the screw tightening work is performed.

[0004]

However, if the visual sensor is temporarily stopped for each work position on the work to correct the work position, the time for calculating the correction movement amount of the work position and the movement for the correction are required. Since the sum of the time and the correction time is added to the working time, the working efficiency cannot be improved. An object of the present invention is to provide a control method for a robot with a visual sensor, which eliminates the above-mentioned drawbacks of the prior art, improves the work efficiency of the robot with a visual sensor, and facilitates the handling of teaching and the like. .

[0005]

In order to achieve the above object, the present invention guides a work position on a work to a position where a visual sensor passes ahead of a work tool,
In addition, by controlling the movement path of the robot head, the visual sensor passes over the work position, the peripheral part of the work position is photographed without stopping while moving, and the gravity center position of the screw hole and the photographing position data calculated from the image data The actual work position is obtained from the attachment position relationship data of the work tool and the visual sensor, and the work tool is moved to the actual work position for work. Further, according to the present invention, teaching work for guiding the sky above the work position on the work to a position where the visual sensor passes ahead of the work tool is omitted, and the current position of the robot head and the next work are eliminated. When the moving path of the head portion is generated from the positional relationship of the positions, the turning arrangement path of the visual sensor is automatically generated.

[0006]

According to the control method described above, the position of the actual work position is corrected within the movement time of the work tool, and the conventional correction time can be omitted. Therefore, the work efficiency can be improved. Further, it is possible to save the teaching work for causing the visual sensor to precede the work tool and to make the handling extremely easy.

[0007]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described with reference to the accompanying drawings. FIG. 1 is a plan view of a screw-fastening robot with a visual sensor (hereinafter simply referred to as a robot) when returning to the machine origin O. A work tool 2 for performing work, a visual sensor 1 attached to a position of a radius R around the axis of the work tool 2 for photographing the work position peripheral portion on the work 3, and a visual sensor for rotationally positioning the visual sensor 1. Motor 4
And a visual sensor encoder 10 directly connected to the visual sensor motor 4. Work 3 has two screw tightening work positions, P1 and P2 represent teaching work positions, and P
It is assumed that 1'and P2 'represent actual work positions.

FIG. 2 shows a block diagram of the visual correction positioning device 11 for controlling the robot 5. The visual correction positioning device 11 includes a CPU 12 that controls the entire visual correction positioning device 11 and a ROM 1 that stores a control program.
4, a keyboard 16 operated by an operator when teaching the position of the robot 5, etc., a RAM 15 for storing teaching data etc. inputted from the keyboard 16, and a photographing command signal 2
3 and a shooting command output unit 17, and a shooting command signal 23
The A / D conversion circuit 18 for converting the analog video signal output from the visual sensor 1 which receives the image upon receiving the digital image into a digital value, the image memory 13 for storing the A / D converted digital image data, and the visual The visual sensor rotational positioning control unit 19 that drives the sensor motor 4 and feeds back the pulse output from the visual sensor encoder 10 to control the rotational positioning of the visual sensor 1.
And an XY table positioning control unit 20 that drives the X-axis motor 6 and the Y-axis motor 7 to move the robot head unit from the current position to the next work position, for example, on the shortest path, that is, a linear path to perform positioning control. , X-axis motor 6
An up / down counter (hereinafter simply referred to as a counter) 2 that counts the number of pulses output from an encoder 8 directly connected to the encoder 8 and an encoder 9 directly connected to the Y-axis motor 7.
1, a shooting position data holding unit 22 that stores and holds the current position of the visual sensor 1 at the moment of receiving the shooting command signal 23 and rotation angle data of the visual sensor 1, and an interrupt signal generation unit 27 that gives an interrupt signal to the CPU 12. , A work tool control unit 28 for controlling the work of the work tool 2, and the like.

FIG. 4 shows a photographed image and a photographing position. 24 is a photographing field of view of the visual sensor 1, 25 is a screw hole which is an object to be recognized, and 26 is a photographing position allowable range. This will be described with reference to the operation flowchart shown in FIG. The operator operates the keyboard 16 according to a predetermined work method, returns the robot 5 to the machine origin O, moves the work tool 2 to a work position on the work 3, and inputs a program such as position teaching and screw tightening work. To do. As shown in FIG. 1, when the robot 5 returns to the mechanical origin O, the visual sensor 1 performs an origin return operation to a position parallel to the X axis, for example.

After completing the program input work, the CPU 12
The teaching work positions P1 and P2 are stored in the RAM 15, the rotation angle θ1 for orienting the visual sensor 1 toward the teaching work position P1 is calculated, and the rotation angle data θ1 is output to the visual sensor rotational positioning control unit 19. The visual sensor rotation positioning control unit 19 drives the visual sensor motor 4 so that the visual sensor 1 passes over the teaching work position P1 in advance of the work tool 2, that is, forms a θ1 angle with the X axis. The visual sensor 1 is rotationally positioned. Also CPU1
2 calculates the photographing position SP1 when the visual sensor 1 comes above the teaching work position P1 from the teaching work position P1, the rotation angle data θ1 and the radius R and stores it in the RAM 15.

Subsequently, the worker inputs a work start command to the robot 5 from the keyboard 16. Upon receiving the work start command, the visual correction positioning device 11 drives the X-axis motor 6 and the Y-axis motor 7 to move the robot head unit from the machine origin position O to the teaching work position P1 (100). ). The XY table positioning control unit 20 controls the robot head unit so that the robot head unit moves along a straight path while feeding back the current position data of the robot head unit. CPU 12
When the current position data of the visual sensor 1 fed back from the photographing position SP1 and the X-axis encoder 8 and the Y-axis encoder 9 falls within a preset photographing position allowable range 26, that is, above the work position on the work 3. When the visual sensor 1 arrives (101), a shooting command is sent to the shooting command output unit 17. The shooting command output unit 17 outputs the shooting command signal 23 to the shooting position data holding unit 22 and the visual sensor 1. The photographing position data holding unit 22 instantly holds the current position (XP1, YP1) of the visual sensor 1 and the rotation angle data θ1 of the visual sensor 1 (102), and the visual sensor 1 is located at the periphery of the screw pilot hole 25 at the working position. Shooting is performed (103), and the video information output from the visual sensor 1 is stored in the image memory 13 via the A / D conversion circuit 18. At this time, the robot head unit continues to move to position the work tool 2 at the teaching work position P1. The CPU 12 calculates the screw hole hole center of gravity position (X1, Y1) in the XY coordinate system from the image data of the image memory 13 as shown in FIG. 4, and stores this screw hole hole center of gravity position (X1, Y1) and photographing position data. Part 22
From the photographing position data (XP1, YP1) and the rotation angle data θ1 of the visual sensor 1, the screw tightening work target position in the XY coordinate system, that is, the corrected position P1 ′ is calculated (104), and the corrected position P1 ′ is controlled by the XY table positioning control. Send to department 20. The XY table positioning control unit 20 changes the positioning target position from P1 to P1 ′ to position the work tool 2 (105), and the work tool control unit 28 controls the work tool 2 to execute the screw tightening work (106).

On the other hand, the CPU 12 performs time-division control from the interrupt signal of the interrupt signal generator 27 after the photographing of the visual sensor 1 is completed in parallel with the above work, and directs the visual sensor 1 to the next teaching work position P2. Therefore, the rotation angle data θ2 of the visual sensor 1 is calculated from the corrected position P1 ′ and the teaching work position P2.
(107), and the visual sensor rotation positioning control unit 1 is calculated.
The rotation angle data θ2 is output to 9. The visual sensor rotational positioning control unit 19 drives the visual sensor motor 4 so as to form a θ2 angle with the X axis to perform rotational positioning of the visual sensor 1 (108). Further, the CPU 12 calculates the photographing position SP2 when the visual sensor 1 comes above the teaching work position P2 on the work 3 from the correction position P1 ′, the teaching work position P2, the rotation angle data θ2 and the radius R and stores it in the RAM 15. (109). The above calculation of the rotation angle θ2 of the visual sensor 1, the rotational angle θ2 of the visual sensor 1, and the photographing position SP2 is performed by the end of the screwing operation of the correction position P1 ′.

When the screw tightening work at the correction position P1 'is completed, the visual correction positioning device 11 moves the robot head portion from the screw tightening work position P1' to the next teaching work position P2 for the X-axis motor 6 and the Y-axis. The motor 7 is driven to start the movement. As described above, the robot head unit moves along a straight path, and an image is taken when the visual sensor 1 comes above the teaching work position P2 on the work 3. The CPU 12 obtains the screw hole hole center of gravity position from the photographed image data, calculates the correction position P2 ′ expressed in the XY coordinate system, and sends the correction position P2 ′ to the XY table positioning control unit 20. The XY table positioning control unit 20 sets the positioning target position of the work tool 2 to P
2 is changed to P2 ′, the work tool 2 is positioned, and the work tool 2 controlled by the work tool control unit 28 performs the screw tightening work.

Although the robot 5 is an orthogonal robot in the above embodiment, it is obvious that the idea of the present invention can be applied to a horizontal articulated robot and other structures. In the above embodiment, the rotation angle θ2 of the visual sensor 1 is calculated, the rotation angle θ2 of the visual sensor 1 is rotationally positioned,
The calculation of the photographing position SP2 is performed by the time when the screw tightening work of the correction position P1 ′ is completed, but it is also possible to perform the calculation on the way to the next teaching work position P2 as shown in the operation flowchart of FIG.

[0015]

As described above, according to the present invention, the position where the visual sensor passes over the work position on the work ahead of the work tool and the movement path of the robot head are controlled to move the work. It is a position correction method that takes a picture around the work position without stopping, finds the actual work position from the image data and the shooting position data, and moves the work tool to the actual work position, so it improves the work efficiency. You can Further, since there is no need to teach the arrangement of the visual sensor, there is an effect that the handling becomes easy.

[Brief description of drawings]

FIG. 1 is a plan view of a screw tightening robot with a visual sensor.

FIG. 2 is a block diagram of a visual correction positioning device.

FIG. 3 is a flowchart for explaining the operation.

FIG. 4 is a plan view showing a photographed image and a photographing position.

FIG. 5 is a flowchart for explaining the operation.

[Explanation of symbols]

1 is a visual sensor, 2 is a work tool, 3 is a work, 4 is a motor for a visual sensor, 5 is a screw fastening robot with a visual sensor, 11 is a visual correction positioning device, 12 is a CPU, 17
Is a photographing command output unit, 19 is a visual sensor rotational positioning control unit, 22 is a photographing position data holding unit, 23 is a photographing command signal, and 25 is a screw pilot hole.

Claims (2)

[Claims] 1. An actual work position, in which a head includes a work tool and a visual sensor, and the visual sensor detects a position of a recognition target object in a work position peripheral portion on a work having at least one work position. In a robot with a visual sensor that obtains data and moves the work tool according to the work position data to perform work, at a position where the visual sensor passes above the work position on the work ahead of the work tool. The visual sensor motor is driven to dispose the visual sensor, and the periphery of the work position is photographed without stopping while moving, the position of the recognition target detected from the image data, the photographing position data, and the rotation angle of the visual sensor. The control of a robot with a visual sensor is characterized in that an actual work position is obtained from data and a work tool is moved to the actual work position. Method. 2. The rotation angle data of the visual sensor is calculated from the positional relationship between the current position of the head portion and the next working position, and the visual sensor motor is driven according to the rotation angle data to arrange the visual sensor. The method for controlling a robot with a visual sensor according to claim 1, wherein the teaching operation of the rotation angle of the visual sensor is omitted.