JPH0647670Y2 - Robot controller - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial application] The present invention relates to a robot controller, and more specifically, a robot that detects the posture of an object held by the robot and changes the posture of the object to a predetermined posture. Regarding the control device.

[Prior Art] In recent years, various industrial robots and robot control devices have become popular with the progress of factory automation. As a control device for such an industrial robot, a robot that detects an image of an object of work, determines the state of the object by processing the image, and controls the operation of the robot based on the determination result. Control devices have been developed. Conventionally, such a robot control device recognizes an object as a two-dimensional image, detects the position on the plane where the object exists, and drives the robot, for example, based on the detection result to drive the object. Control was performed such that an object was gripped or moved.

[Problems to be Solved by the Invention] However, in such a conventional robot control device, it may not be possible to accurately determine the posture of an object. For example, when one object is grasped and moved from a plurality of objects randomly piled up on a pallet, a two-dimensional image obtained by imaging the piled up objects from one direction It is extremely difficult to judge the accurate posture of the target object. That is, although the outlines of the positions and orientations of a plurality of objects can be recognized, the accurate orientations of individual objects cannot be determined.

For example, when the shafts with flanges having a small difference between the flange outer diameter and the shaft outer diameter are piled up, it may not be possible to determine which is the flange side.

For this reason, although the robot can grasp and take out individual objects, their postures are inaccurate.

Further, even if such a state in which a plurality of objects are piled up is imaged from a plurality of directions, it is still impossible to accurately determine the posture of each object.

As a result, the individual objects taken out by the robot cannot be supplied to the next process in an accurate posture, which hinders work in the next process.

Further, it is desired that the work of individually feeding the loaded objects to the next process be performed promptly in order to improve the work efficiency.

The present invention solves the above-mentioned problems, and even when a plurality of objects are stacked, each object can be supplied in a proper posture to the supply target part in the next process and the like. An object of the present invention is to provide a robot control device that is simple and capable of rapid processing.

[Means for Solving the Problems] In order to achieve the above object, the present invention has a configuration as shown in FIG. FIG. 1 is a block diagram showing the basic concept of the present invention.

As shown in FIG. 1, the present invention comprises a first image detecting means b for detecting images of a plurality of objects stacked on a mounting portion from a first direction, and a first image detecting means b. State determination means d for determining the state of at least the position and inclination of one of the objects c among the objects a based on the detection result, and the object determined based on the determination result of the state determination means d Control means f for driving the robot e which grips the object c and moves the object c to supply the object c to the supply target portion.
In a robot control device including: an image of the object c gripped by the robot at a predetermined position on the movement path from the placement unit to the supply target unit;
The second detecting from the second direction which is non-parallel to the first direction
Image detection means g, the attitude determination means h for determining the attitude of the object c at the predetermined position based on the detection result by the second image detection means g, and the determination result of the attitude determination means h. Robot e
And a posture changing means i for changing the posture of the object c to a preset regular posture, and a robot controller characterized by the following.

[Operation] Next, the operation of the present invention will be described with reference to FIG.

In the robot controller of the present invention, the first image detecting means b detects the images of the plurality of objects a stacked on the mounting portion from the first direction, and the state determining means d determines the detection result based on the detection result. The state of at least the position and inclination of one object c is determined. Based on the determination result, the control means f drives the robot e to grip and move the object c and supply the object c to the supply target portion, but before supplying the object c,
It operates as follows.

That is, the image of the target object c at a predetermined position on the path for moving the target object c from the placement section to the supply target section is detected by the second image detection means g from the second direction that is not parallel to the first direction. Then, the posture determination means h determines the posture of the object c based on the detection result. Then, based on this determination result, the posture changing means i drives the robot e to change the target object c to a preset normal posture.

In this case, the posture of only one target object c is determined after gripping, that the target object c has a restricted posture because it is gripped by the robot e, and the target object c stops at a predetermined position. Therefore, the distance between the object c and the second image detecting means g is constantly constant, and the accuracy is not deteriorated due to the distance difference. Therefore, the image detected by the first image detecting means b cannot be read. The missing details can be read from the image detected by the second image detecting means g. Therefore, the posture determination means h can easily determine the posture of the target object based on the image detected by the second image detection means g.

Therefore, the target object c is constantly supplied to the supplied portion in a regular posture.

As described above, each component of the present invention acts to solve the technical problem of the present invention.

[Embodiment] Next, a preferred embodiment of the present invention will be described with reference to the drawings. FIG. 2 is a system configuration diagram of a rough material take-out robot controller according to an embodiment of the present invention.

In FIG. 2, the articulated robot 10 is controlled by a robot controller (hereinafter simply referred to as RC) 12 having an operation panel 11. An ultrasonic sensor 16 for detecting an obstacle is arranged in the vicinity of the hand 14 of the articulated robot 10, and a proximity switch 18 for confirming gripping of a part is further provided in the hand 14 of the articulated robot 10. It is built in.

The direction perpendicular to the plane in which the articulated robot 10 is installed is the Z direction indicated by arrow C in the figure, and two directions orthogonal to each other in the plane orthogonal to the Z direction are indicated by arrow A.
And the Y direction shown by the arrow B. Then, in this embodiment, a three-dimensional space having the above-mentioned three directions as coordinate axes is set, and the following description will be given.

The first camera 22 allows the parts pallet 24 to be detected as an image in the Z-axis direction from the parts 26 consisting of the flange parts stacked in the parts pallet 24 and the shaft part having an outer diameter slightly smaller than the flange parts. It is installed facing down just above. On the other hand, the second camera 28 is arranged so that the component 26a lifted by the hand 14 of the articulated robot 10 and moved to a predetermined position can be detected as an image from the Y-axis direction as an image.
It is installed horizontally in the axial direction. The fields of view of both cameras are shown by a two-dot chain line in FIG. The image data captured by the first camera 22 and the second camera 28 is input to the image processing device 30, and the various data obtained by the image processing is transmitted from the image processing device 30 to the RC 12.
On the other hand, a signal indicating the state of the articulated robot 10 is transmitted from the RC 12 to the image processing device 30. Parts 26b are arranged and arranged on the conveyor 32 by the articulated robot 10.

In the present system configured as described above, the image data detected by the first camera 22 and the second camera 28 is processed by the image processing device 30, and based on the data output by the image processing device 30, the RC12 The articulated robot 10 is controlled to move the parts 26 in the parts pallet 24 onto the conveyor 32.

Next, the configurations of the RC 12 and the image processing device 30 will be described with reference to FIG.

As shown in the figure, the RC 12 includes the operation panel 11, the ultrasonic sensor 16, the proximity switch 18, and the image processing device 30 described later, which are described above.
A central processing unit (hereinafter simply referred to as a CPU) 12a that performs processing for controlling the operation of the articulated robot 10 described above while inputting and calculating respective signals or data input from the control program according to a control program,
Read only memory (hereinafter simply referred to as ROM) 12b, R in which the above control program and initial data are stored in advance
It is provided with a random access memory (hereinafter, simply referred to as RAM) 12c in which various signals input to C12 and data necessary for arithmetic control are temporarily stored.

RC12 is the ultrasonic sensor 16 and the proximity switch 18 described above.
A waveform shaping circuit 12d that shapes the waveform of the output signal of the above and a drive circuit 12e that drives the robot 10 are provided. Then, the signals from the operation panel 11, the ultrasonic sensor 16, the proximity switch 18, and the data from the image processing device 30 are transmitted to the CPU 12a via the input / output port 12f. Further, from the input / output port 12f, the drive circuit 12e of the articulated robot 10 is
To the image processing device 30 and the control signal to the image processing device 30 is output. The RC 12 is provided with a bus line 12g which serves as a passage for control signals and data to the above elements.

On the other hand, the image processing device 30 inputs and analyzes the image data input from the first camera 22 and the second camera 28 described above according to the control program, and performs the processing for controlling the data communication to the RC12 described above. The CPU 30a to be executed, the control program and the initial data are stored in advance.
A ROM 30b and a RAM 30c for temporarily storing image data input to the image input device 30 and data necessary for control are provided.

Further, the image processing device 30 includes buffers 30d and 30e for image signals output from the first camera 22 and the second camera 28,
An A / D converter 30f that converts the image signal that is an analog signal into a digital signal, and a multiplexer 30g that selectively outputs the output of the A / D converter 30f to the CPU 30a are provided. The image data from the first camera 22 and the second camera 28 and the control signal from the RC 12 are input / output port 30.
It is transmitted to the CPU 30a via h. Furthermore, the data to RC12 is output from the input / output port 30h. The image processing device 30 includes a bus line 30i that serves as a passage for control signals and data to each of the above elements.

The processing executed by the RC 12 and the image processing apparatus 30 will be described below with reference to the flowcharts of FIGS. 4 (a) and 4 (b). FIG. 4A is a flowchart showing the component alignment processing executed by the RC 12, and FIG. 4B is a flowchart showing the image processing executed by the image processing device 30. Both processes are RC when the parts alignment work is performed by the operator's instruction or the request of the automatic machine.
12 and the image input device 30 perform data communication with each other so that they are executed in parallel.

After starting RC12, workers and automatic machines select parts alignment work,
When a command is input from the operation panel 11, the parts alignment processing shown in FIG. 4 (a) is executed.

First, in step 100, an image processing start command is output from the RC 12 to the image processing device 30. Based on this command, the image processing device 30 starts execution of the image processing shown in FIG. The RC12 is the arm 15 of the articulated robot 10.
Is retracted out of the field of view of the first camera 22.

Based on the command of step 100 executed by the RC12, the image processing device 30 executes step 400 of the image processing shown in FIG. 4B. Here, the first camera 22 is selected as the image data input device. Then step 41
Proceeding to 0, the component image is detected from the first camera 22. That is, the first camera 22 images the state of the components 26 stacked in the component pallet 24 in the Z-axis direction. In the following step 420, the image processing device 30 scans the image data of the component 26 detected by the first camera 22 in step 410, and recognizes the overall shape of the component 26. Then, the type of the component 26, the position on the XY plane, and the inclination from the basic posture are determined.

On the other hand, the RC 12 executes step 110 shown in FIG. 4A while the image processing device 30 is performing the above-described processing. That is, it waits until the recognition completion signal is transmitted from the image input device 30.

The image processing device 30 executes step 430 following step 420 described above. That is, the recognition completion signal is
Output to. Then, proceeding to step 440, the image processing apparatus 30 sets the position of the component 26 on the XY plane to the plane coordinate value (x, y), and the inclination of the component 26 from the basic posture is the rotation value (θ). Further, the type of the component 26 is output to the RC 12 as a predetermined code.

RC12 is the step 43 executed by the image processing apparatus 30 described above.
In response to the recognition completion signal output of 0, the process proceeds from step 110 to step 120. Then, the image processing device 30 described above
Input the output data of step 440 to be executed in RAM12c
Stored in a predetermined area. Then RC12 goes to step 130
To execute. That is, based on the plane coordinate value (x, y) of the component 26 input in step 120, the arm 15 of the articulated robot 10 is moved right above the component 26 on the XY plane. Further, the tilt of the hand 14 of the articulated robot 10 is adjusted based on the rotation value (θ) of the component 26 input in step 120. In the following step 140, the arm 15 of the articulated robot 10 is lowered at high speed in the Z-axis direction. Then, the process proceeds to step 150, and the output of the ultrasonic sensor 16 described above is detected. Further, in step 160, it is determined whether or not the output of the ultrasonic sensor 16 is turned on. That is, it is determined whether the arm 15 of the articulated robot 10 descends at high speed in the Z-axis direction and approaches the part 26. When the output of the ultrasonic sensor 16 is OFF, it is determined that the arm 15 is not close to the component 26, the process returns to the step 140, and the high speed descent in the Z-axis direction is continued. On the other hand, if the output of the ultrasonic sensor 16 is turned on, the arm 15
And the process proceeds to step 170. Where R
The arm 15 of the articulated robot 10 is lowered at C12 in the Z-axis direction at a low speed. In the following step 180, the output of the proximity switch 18 incorporated in the hand 18 of the articulated robot 10 is detected. Further, in step 190, it is determined whether or not the output of the proximity switch 18 is turned on. That is, the hand 14 of the articulated robot 10 descends at a low speed in the Z-axis direction and the parts 26
It is determined whether or not the user has touched. When the output of the proximity switch 18 is OFF, it is determined that the hand 14 is not in contact with the component 26, the process returns to step 170, and the low speed descent in the Z-axis direction is continued. On the other hand, when the output of the proximity switch 18 is turned on, it is determined that the hand 14 has come into contact with the component 26, and the process proceeds to step 200. Here, RC12 is arm 1
Stop the descent of 5 in the Z-axis direction. And step 210
Then, the hand 14 grips one component 26a of the components 26 in the component pallet 24. In the following step 220, the arm 15 rises in the Z-axis direction while the hand 14 holds the component 26a. Here, the gripped state of the component 26a is constantly monitored based on the output of the proximity switch 18. Then, it proceeds to step 230. Here, the component 2 gripped by the hand 14
The arm 15 is moved so that 6a is within the field of view of the second camera 28 described above. Then proceed to step 240,
The part 26a gripped by the hand 14 is set in advance in the RAM 12
It is determined whether or not it has been positioned at the predetermined position stored in c. And until it is positioned at that position,
The above steps 230 and 240 are repeated. When the movement of the arm 15 is continued and the above positioning is completed, the routine proceeds to step 250, where the RC 12 outputs a fixed position stop signal to the image processing device 30.

On the other hand, the image processing apparatus 30 performs the steps shown in FIG.
At 450, the second camera 28 is selected as the image data input device. Then, the routine proceeds to step 460, where the RC 12 waits until a fixed position stop signal indicating that the component has stopped at the fixed position is output. As described above, when the RC 12 outputs the home position stop signal in step 250, the image processing device 30
Performs step 470. Here, the direction of the part 26a from the second camera 28 is detected. That is, the second camera 28 images the direction of the component 26a held by the hand 14 from the Y-axis direction. In the following Step 480, the above Step 470 is performed.
Then, the image processing device 30 scans the image data of the component 26a detected by the second camera 28 to recognize the orientation of the component 26a. For example, it is determined whether the flange portion of the component 26a is on the right side or the left side with respect to the Y-axis direction. Then, the processing proceeds to step 490, and the image processing device 30 outputs the component direction data indicating whether the component 26a faces rightward or leftward to the RC 12. Then, the image processing ends.

On the other hand, RC12 is the same as step 490 described above by the image processing device 30.
The component direction data output in step 4 is input in step 260 shown in FIG. Then, the process proceeds to step 270, and based on the data input in step 260, it is determined whether or not the component 26a gripped by the hand 14 is facing right. If the part 26a faces rightward, the process proceeds to step 280, where the hand 14 is rotated 90 ° clockwise about the Z axis. Then, the process proceeds to step 300. On the other hand, if the component 26a is facing left, the process proceeds to step 290, where the hand 14 is rotated 90 ° counterclockwise around the Z axis. And step 300
Proceed to. In the following step 300, the part 26a oriented in the predetermined direction in step 280 or 290 described above is moved to the arm 15a.
After being placed at a predetermined position on the conveyor 32 by moving, the process goes to NEXT and the RC 12 ends the parts alignment process. Then, the control shifts to the next process selected by the operator.

In the present embodiment, the object a is the component 26, the first image detecting means b is the first camera 22, the image processing apparatus 30, and the processing executed by the image processing apparatus 30 (400, 410).
The object c is the component 26a, the state determination means d is the image processing apparatus 30 and the processing (420, 430, 440) executed by the image processing apparatus 30, and the robot e is the articulated robot 10.
In addition, the control means f controls the RC12 and the processing (100, 110, 120, 130, 140, 150, 160, 170, 180, 190, 200, 21) executed by the RC12.
0,220,230,240,250) respectively. Also, the second
The image detecting means g of the second camera 28, the image processing apparatus 30, and the processing executed by the image processing apparatus 30 (450, 460, 47).
0), the attitude determination means h corresponds to the image processing apparatus 30 and the processing (480,490) executed by the image processing apparatus 30, and the attitude changing means i corresponds to the RC12 and the processing (260,270,280,290) executed by the RC12. .

As described above, in the present embodiment, the position, the inclination, and the type on the XY plane of the components 26 stacked in the component pallet 24 are detected and recognized by the first camera 22 and the image processing device 30. RC12 articulated robot based on recognition 10
Is controlled to grasp one component 26a and move it to a predetermined position. Then, the direction of the component 26a is detected and recognized by the second camera 28 and the image processing device 30, and the RC1 is detected based on the recognition.
2 controls the articulated robot 10 to align the parts 26a in a predetermined direction and arrange them on the conveyor 32. For this reason,
Even if it is difficult to determine the orientation of the component 26a based on the image data detected by the first camera 22, the orientation can be accurately determined by recognizing the component 26a based on the image data detected by the second camera 28. it can. This is because the component 26a imaged by the second camera 28 is grasped by the hand 14 in a state in which the direction is restricted to either the right or the left, so that the determination of the direction becomes easy. . Moreover, since the component 26a is positioned at the predetermined position,
There is also no risk of a reduction in accuracy due to perspective.

In addition, before gripping, the parts are based only on the image of the first camera 22.
Since the configuration is such that the position of 26 is recognized and the orientation of the component 26a is determined only from the image of the second camera 28 after gripping, the device is simplified in terms of both hardware and software, and processing can be executed quickly.

Further, since the first camera 22 images a plurality of components 26, a wide range of images is obtained, but the second camera 28 captures only the component 26a that is once held, so that it is possible to make a detailed posture determination, and the determination accuracy increases. . Moreover, once the component 26a is gripped, the orientation of the component 26a can be detected by imaging in only one direction.

Further, with the above effect, the accuracy of the component supply position is increased,
It is possible to unify the orientations of parts of arbitrary shapes stacked in bulk and arrange them in alignment. As a result, the efficiency of the parts alignment work is improved, and the reliability of the work is increased.

Further, in the present embodiment, the first camera and the second camera are installed in the directions orthogonal to each other, so that the position coordinates of the component and the orientation of the component can be determined by a simple calculation.

In this embodiment, two cameras, a first camera and a second camera, are used. However, for example, one camera and a device for moving and positioning the camera to two positions that are not parallel to each other are used. Even so, the effect of the present invention can be obtained.

Further, in this embodiment, in order to unify the directions of the parts, Z
Rotated 90 ° left and right about the axis, but 180
Even if it is rotated to the left or right by an angle forming a supplementary angle of °, the same effect as this embodiment can be obtained.

Furthermore, in the present embodiment, the orientations of the components are all unified and aligned, but the orientations of the components may be arranged in a staggered manner, or the orientations of the components may be changed for each predetermined number.

In the present embodiment, the operation of changing the orientation of the component was performed at a predetermined stop position, but for example, the operation is performed on the component pallet 24, or while moving from the component pallet 24 onto the conveyor 32, or It may be performed immediately before the placement on the conveyor 32.

Although the embodiments of the present invention have been described above, the present invention is not limited to such embodiments, and it is needless to say that the present invention can be implemented in various modes without departing from the scope of the present invention. .

[Effects of the Invention] As described in detail above, according to the robot control device of the present invention, first, the position and inclination of the object are recognized based on the image of the object from the first direction, and then the robot 1 The posture of the object can be accurately recognized only by gripping one object and then detecting the image of the object from the second direction. Therefore, the plurality of loaded objects can be individually supplied to the supply target portion in a preset normal posture. Further, the device is simplified in terms of both hardware and software, and the processing can be executed quickly.

[Brief description of drawings]

FIG. 1 is a block diagram showing the basic concept of the present invention, FIG. 2 is a system block diagram of a rough material take-out robot control device according to an embodiment of the present invention, and FIG. 3 is the robot controller (RC) and image processing thereof. 4 is a block diagram for explaining the configuration of the apparatus, FIG. 4 (a) is a flow chart showing the processing executed by the RC in one embodiment of the present invention, and FIG. 4 (b) is image processing in one embodiment of the present invention. It is a flow chart which shows processing performed by a device. a ... Object, b ... First image detecting means c ... Object, d ... State determining means e ... Robot, f ... Control means g ... Second image detecting means, h ... Posture determination means i ... Posture change means 10 ... Articulated robot 12 ... Robot controller (RC) 22 ... First camera 26, 26a, 26b ... Parts 28 ... Second camera 30 ... Image processing device

Continuation of the front page (72) Inventor Hiroyuki Oishi 1 Toyota Town, Toyota City, Aichi Prefecture Toyota Motor Corporation (56) Reference JP-A-53-127988 (JP, A)

Claims (1)

[Scope of utility model registration request] 1. A first image detecting means for detecting an image of a plurality of objects stacked on a placing part from a first direction, and the object based on a detection result by the first image detecting means. Of at least one of the objects, the state determining means for determining the state of the position and the inclination, the object determined based on the determination result of the state determining means, and the object is moved. A robot controller for driving a robot to supply the supply target to the supply target, and an object grasped by the robot at a predetermined position on the movement path from the placement unit to the supply target. Second image detecting means for detecting an image from a second direction which is non-parallel to the first direction, and a posture of the object at the predetermined position is determined based on a detection result by the second image detecting means. Posture determination hand And a posture changing unit that drives the robot based on the determination result of the posture determining unit and changes the posture of the object to a preset normal posture. .