JPH06206186A - Calibration of handling precision of horizontal articulated type robot equipped with visual sensor - Google Patents

Abstract

PURPOSE:To carry out the correct handling work of a robot from the position data of a workpiece which is obtained by a visual sensor. CONSTITUTION:The handling work for a workpiece 13 is carried out, by carrying out the correction calculation fat the position data of a visual sensor according to the correction value table, by providing the correction value table having the deflection quantity generated between the handling position A of the workpiece 13 as working object and the tool shaft 8 of a robot and the offset quantity between the hand standard position B of the robot and the tool shaft 8 of the robot, as parameter, for each image pick-up position of a camera means.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for calibrating position data when a robot performs a handling operation using position data obtained through a visual sensor.

[0002]

2. Description of the Related Art Conventionally, a position data input method by teaching has been used as a position command method for an industrial robot.
On the other hand, in recent years, there has been an increasing demand for a method of operating a robot using position data created offline or position data obtained via a visual sensor. But,
Generally, for industrial robots, the positioning reproducibility of the robot itself is satisfactory, but the relative positioning accuracy between the robot and the outside world remains a problem.

As a reference example of the prior art, a teaching method of a robot coordinate system for a visual sensor shown in FIG. 5 will be given. In this example, the camera means 7 is attached to the tip of the second arm 6 of the horizontal articulated robot R so that the visual field direction is vertical to the robot work surface, and the imaging position is moved to an arbitrary point within the robot operation range. It is possible. A plate 14 for teaching a visual sensor is installed on the work surface, and the plate 14 has three positioning holes 15 that are generally spaced from each other.
a, 15b, and 15c are provided. The camera means 7 can simultaneously accommodate these three positioning holes in the field of view, and the camera image position calculation means can calculate each position as position data of the camera coordinate system. Further, the tool shaft 8 has three positioning holes 15a, 1
It can be inserted into 5b and 15c, and the position data of the robot coordinate system can be grasped.

First, the horizontal articulated robot R is moved to a position where the camera means 7 can image the three positioning holes 15a, 15b, and 15c at the same time, and each positioning hole is imaged at the imaging position to determine the position of the center of gravity. The position data in the camera coordinate system is calculated by the camera image position calculation means. Since the distance between the three positioning holes is generally known, the size of each pixel in the camera coordinate system can be grasped at this stage. Next, the tool shaft 8 is inserted into each positioning hole to grasp the position data of each positioning hole in the robot coordinate system. From the position information of the three points indicated by each coordinate system,
It is possible to calculate the distance L _D between the center position D of the camera means 7 and the tool axis 8 and the tilt angle θ _D with respect to the second arm 6, and each joint angle and each arm length of the horizontal articulated robot R, and It is stated that the position data of the camera coordinate system at an arbitrary imaging position can be converted into the position data of the robot coordinate system by using the parameter of the mounting position of the camera means 7.

[0005]

In the conversion method from the camera coordinate system to the robot coordinate system, the robot coordinate system itself is distorted unless the arm lengths and joint angles of the robot are known in advance. However, the attachment dimension of the camera means to the robot arm cannot be calculated accurately. Generally, it is easy to grasp each joint angle of the robot, but it is difficult to accurately determine each arm length, and the arm length changes due to thermal expansion or the like. Regarding the camera mounting dimensions for the robot, the mounting angle of the camera means with respect to the arm can be accurately determined, but it is difficult to accurately determine the mounting distance with respect to the tool axis. Therefore, when converting the position data of the camera coordinate system obtained by the camera image position calculating means into the position data of the robot coordinate system, the error is mainly caused by the error of each arm length of the robot and the attachment distance of the camera means to the tool axis. Will occur. Also, since the posture of the work whose position data is recognized by the visual sensor is arbitrary, the robot rotates the hand attached to the tool axis to an appropriate angle to pick up and place the work. If you do not know the offset amount from the axis, you cannot perform accurate handling work.

In order to solve these problems, an object of the present invention is to accurately handle a work using a camera means attached to an arm of a horizontal articulated robot and position data obtained from a camera image position calculating means. To provide a method of calibrating position data obtained from a visual sensor in order to do so.

[0007]

In order to solve the above problems, according to the present invention, when the tool axis of the robot is moved to the work recognition position obtained from the visual sensor, it is preset in the work area of the robot. For each camera image capture position, there is a correction value table that uses the amount of deviation between the handling position of the work to be worked and the tool axis of the robot and the offset amount between the robot hand reference position and the robot tool axis as correction value parameters. Then, the robot operates while correcting and calculating the position data obtained from the visual sensor with the correction value parameter of the correction value table corresponding to the camera image pickup position, thereby realizing an accurate handling work using the visual sensor.

[0008]

When there is an error in the length of each arm of the robot held by the robot controller or the camera mounting dimension, if the image pickup position of the camera means changes at any position in the robot work area, the work recognition obtained by the camera image position calculation means. When the tool axis of the robot is moved to the position, the amount of deviation that occurs between the handling position of the work to be processed and the tool axis of the robot varies depending on the camera image capturing position. A correction value table for correcting is required. Further, since the posture of the work whose position data is recognized by the camera image position calculation means is arbitrary, the robot rotates the hand attached to the tool shaft at an appropriate angle to pick and place the work. Therefore, the robot that handles the work with the position data obtained from the camera image position calculation means is
For each imaging position of the predetermined camera means, the deviation amount between the handling position of the work object and the tool axis of the robot, and the offset amount between the robot hand reference position and the robot tool axis are used as correction value parameters. It is possible to move the hand reference position of the robot to the handling position of the work to be worked by correcting and calculating the position data of the visual sensor using the individual correction value table.

[0009]

Embodiments of the present invention will now be described in detail with reference to the drawings. FIG. 1 shows an appearance of a horizontal articulated robot R with a visual sensor, which is cited in this description. Here, a first arm 4 that can swivel around the first swivel axis 3 in the direction of the work surface 11 is provided on the base 2, and the first arm 4
Is provided with a second arm 6 which can be swung in the direction of the work surface 11 around the second swivel shaft 5 located at the tip of the tool shaft 8 which is vertically movable at the end of the second arm 6 and can be swung in the direction of the work surface 11
An embodiment will be described by taking a horizontal articulated robot R having A hand 9 for handling a work by a suction mechanism is attached to the tip of the tool shaft 8. The tip of the suction nozzle of this hand 9 is processed into a taper shape so that an operator can hold the hand 9 with respect to the suction position of the work. It is easy to teach the tip of. At the tip of the second arm 6 of this robot, a camera means 7 for capturing an image is attached so that the visual field direction is vertical to the work surface 11. The robot controller 1 includes a robot motion control means and a camera image position calculation means. A teaching device 10 that performs a teaching operation is attached to the robot controller 1.

FIG. 2 is a diagram showing the handling work performed by the horizontal articulated robot R described in FIG. In the work shown in this figure, parts feeders 12a, 12b, and 12c corresponding to the visual robot are installed in the work area. Each parts feeder does not align the parts at a predetermined position, but feeds the work into each imaging region indicated by a rectangle in the drawing in any posture. Robots
The camera means 7 is located at the camera image capturing point defined on each image capturing area.
Is moved to image the work. To the operation control means and the camera image position control means of the robot controller 1, parameters such as the arm lengths of the robot, the joint angles, the camera mounting dimensions, and the pixel unit dimensions, which are obtained in advance, are input in advance. The camera image position control means, the position data in the camera coordinate system of the imaged work,
By converting the position data of the robot coordinate system using these parameters and passing them to the robot operation control means, the robot can recognize the position data of any work in the imaging area of each parts feeder and perform the work of handling the work. It has become like.

When the actual arm lengths of the robot and the camera mounting distance with respect to the tool axis 8 have errors with the parameters held by the robot controller 1, the camera coordinate system is converted to the robot coordinate system as shown in FIG. Although the position data of the camera image position calculating means is expressed by xy coordinates, it has a certain offset amount with respect to the robot coordinate system (XY coordinates) of the robot operation control means. In such a case, the position data of the work 13 in the imaging area of the parts feeder 12a is obtained from the camera image position calculation means via the camera means 7, and the robot operation control means sets the tool axis 8 to the handling position of the work 13. When operated, the state of the model shown in FIG. 4 is obtained. Like this
The tool axis 8 moves to the position of the point D with respect to the point A which is the handling position of the work 13. Also, hand 9
When the reference position of 1 has an offset amount with respect to the tool axis 8 of the robot, the reference position of the robot hand 9 is displaced to the point B when the angle of the tool axis 8 is 0 degree.

The image pickup position of the parts feeder 12a of the camera means 7 is taught as a fixed point which is not changed during the robot work, and the image pickup area of the parts feeder 12a is a limited area with respect to the operation range of the robot. .
Therefore, the amount of deviation generated between the point A indicating the handling position of the work shown in FIG. 4 and the point D indicating the position of the tool axis 8 of the robot is a constant value and is represented by one approximate value for the imaging area. be able to. From this, work 13
The correction value parameter for accurately handling the work at an arbitrary position in an arbitrary posture in the imaging area of the parts feeder 12a by measuring the amount of deviation of the point D and the point B from the point A indicating the handling position. Can be asked.

First, the position of the point B shown in FIG. 4 with respect to the point A is obtained. Point A indicating the handling position of the work 13
Then, the reference position of the hand 9 at the point B is moved by using the teaching device 10 while keeping the tool angle at 0 degree. X direction at that time,
The position of the point B with respect to the point A on the XY plane can be grasped from the amount of movement in the Y direction. Further, if the tool shaft 8 of the robot is rotated 180 degrees after the reference position of the hand 9 is returned to the position of the point B, the reference position of the hand 9 comes to the position indicated by the point C.
Similarly, from the position of this point C, the reference position of the hand 9 is kept at the tool angle of 180 degrees by using the teaching device 10.
To the X and Y directions, and then XY
The position of the point C with respect to the point A on the plane can be grasped.

Point D indicating the position of the tool axis 8 of the robot
The position with respect to the point A indicating the handling position of the workpiece 13 is obtained as the midpoint between the points B and C. Therefore,
When the coordinates of the point A are (x _A , y _A ), the coordinates of the point B are (x _B , y _B ), and the coordinates of the point C are (x _C , y _C ), the center of the tool axis 8 is Point D (coordinates (x _D ,
y _D)) and the distance between the point A indicating the handling position of the workpiece 13 L _DA, and angle theta _DA with respect to the X-axis of the vector DA
Is calculated from equation 1.

[0015]

[Equation 1]

Further, when the tool axis 8 has an angle of 0 degree, the distance L _BD between the point B indicating the center of the hand 9 and the point D indicating the center of the tool axis 8 and the angle θ _{BD of the} vector BD with respect to the X axis.
Is calculated from the equation 2.

[0017]

[Equation 2]

Let W (x, y) be the suction point of the workpiece having an arbitrary posture in the image pickup area of the parts feeder 12a obtained through the camera means 7 and the camera image position calculation means.
Assuming that the angle of the tool shaft 8 at the time of sucking the work is θ, the correction calculation formula for moving the reference position of the hand 9 to the suction point W of the work can be obtained by Expression 3.

[0019]

[Equation 3]

The parameters used in the above correction arithmetic expression are
By obtaining for each imaging position by the method described above, the robot can accurately perform the work of handling any work in a predetermined area according to the work handling position obtained from the visual sensor. . Further, when working while exchanging hands with respect to a plurality of works, similarly accurate handling work can be realized by obtaining the parameters used in the correction calculation formula for each work.

In the above example, the camera means is attached to the tip of the arm. However, as long as the height and angle of the camera means with respect to the work surface do not change when the robot is operated, the mounting position of the camera means Without being limited, the same correction calculation can be performed on the position data obtained from the visual sensor. Further, the hand attached to the tool shaft is not limited by this embodiment as long as the relationship between the reference position of the hand and the handling position of the work to be worked can be grasped.

[0022]

When the camera means is used so as to be movable by the robot operation, the imageable area can be widened, but when the position data of the visual sensor is converted from the camera coordinate system to the robot coordinate system, the robot's coordinate system is changed. If the length of each arm and the camera mounting distance to the tool axis cannot be accurately grasped, the accuracy of the position data cannot be guaranteed. However, it is difficult to guarantee the accuracy of the position information of the visual sensor in the entire robot movable range by accurately grasping those parameters. Therefore, in the handling accuracy calibration method according to the present invention, the position data of the visual sensor for the predetermined imaging area is corrected by taking into account the offset amount of the hand with respect to the tool axis and performing a correction calculation, thereby It is possible to guarantee the accuracy of the handling work for any work.

[Brief description of drawings]

FIG. 1 is an external view of a horizontal articulated robot with a visual sensor cited in an embodiment of the present invention.

FIG. 2 is a configuration diagram of handling work performed by a horizontal articulated robot with a visual sensor, which is cited in an embodiment of the present invention.

FIG. 3 is a model showing the deviation of the coordinate system of the visual sensor cited in the example of the present invention.

FIG. 4 is a model showing a correction method for position data of a visual sensor according to the present invention.

FIG. 5 is a configuration diagram for teaching a robot coordinate system to a visual sensor of a conventional art example.

[Explanation of symbols]

 1 Robot Controller 6 Second Arm 7 Camera Means 8 Tool Axis 9 Hand 10 Teaching Device 13 Workpiece

Claims (2)

[Claims] 1. A robot arm is provided with a camera means in a vertical direction on a work surface, the work position is recognized through a camera image position calculation means, and the work axis is maintained in a vertical posture with respect to the work surface. In the horizontal articulated robot that performs the handling work of, when the tool axis of the robot is moved to the position data of the work obtained from the camera image position calculation means, the deviation amount between the handling position of the work and the tool axis of the robot , A correction value table for correcting the offset amount between the robot hand reference position and the robot tool axis is set for each camera image capturing position,
A method for calibrating the handling accuracy of a horizontal articulated robot with a visual sensor, characterized in that the robot is operated while correcting and calculating the position data of the work obtained from the camera image position calculation means using a correction value table for each imaging position. 2. The handling accuracy calibration method according to claim 1, wherein between a handling position of a work to be worked and a tool axis of a robot with respect to an arbitrary camera image capturing position necessary for creating a correction value table. The amount of deviation that occurs,
A method for calibrating the handling accuracy of a horizontal articulated robot with a visual sensor, characterized in that the offset amount between the robot hand reference position and the robot tool axis is measured simultaneously by the teaching work of the hand reference position for the work to be worked.