JPH012883A - robot control device - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION Summary of the Invention By configuring a control system using target values and measured values expressed in a camera coordinate system, high-speed processing and high precision are made possible.

BACKGROUND OF THE INVENTION Technical Field The second invention relates to a robot throat control device that is particularly applicable to a robot system equipped with a visual device in which the position of a workpiece to be gripped is not necessarily certain.

Prior art and its problems In order to accurately set the position and orientation of a workpiece to be gripped by a robot, expensive jigs and conveyance devices are required. Therefore, in order to reduce the cost of equipment, we did not require precision in setting the position of the workpiece.Instead, a visual device was attached to the robot's hand, and this visual device measured the position and orientation of the workpiece, and that information was also used. In general, a method is adopted in which the hand of the robot is positioned on the workpiece.

The conventional method for positioning a robot's hand is to determine the position of the workpiece using the visual device's own coordinate system, convert this to the robot's base coordinate system, and then convert that data into the robot's joint angles. Ta. This joint angle becomes the robot's motion target value.

The robot is controlled so that its hand approaches the target value.

In such a conventional method, there are many calculations of coordinate conversion, and this calculation is complicated, so it is difficult to speed up the operation and high accuracy cannot necessarily be obtained.

SUMMARY OF THE INVENTION OBJECT OF THE INVENTION The purpose of this invention is to achieve high-speed operation of a robot and high viscosity positioning.

Structure and Effects of the Invention The robot control device according to the present invention includes a visual device provided to maintain a predetermined positional relationship with a hand effector held on an arm member of the robot, and a workpiece image grasped by the visual device. location and! - a first calculation means for calculating, in the coordinate system of the visual device, a quantity related to the deviation from the target position determined by the predetermined positional relationship; Driving the arm member so that the position of the two-handed effector approaches the 1-1 mark position according to the predetermined distance obtained by the second calculation means for converting into a predetermined amount in the coordinate system and the twenty-first calculation means. .

It is characterized by comprising means for controlling.

This invention can be applied after the image of the workpiece has entered the visual field of the visual device, and conventional control methods can be used when the workpiece is in a range that cannot be grasped by the visual device. A predetermined position (eg, a central position) of the hand effector is within the field of view of the vision device.

According to this invention, after the workpiece enters the field of view of the camera of the visual device, the 1" target position of the control system is fixed at a constant coordinate position within the camera coordinate system, and the workpiece also represented by the camera coordinate system is fixed. Control is performed using the image position as feedback information to eliminate deviations between them. Since both the target position and the feedback information are expressed in the camera coordinate system, the deviation calculation by the first calculation means is extremely simple, and the conversion calculation of the coordinate system can be performed by the second calculation means. Moreover, this calculation is also relatively simple.

Therefore,: (The number of complicated coordinate conversion calculations is reduced, high-speed operation becomes possible, and a decrease in accuracy due to conversion calculations can be prevented, making it possible to perform high-precision positioning.

DESCRIPTION OF THE EMBODIMENTS FIG. 1 shows an example of a typical planar working robot configured with rotary joints. The robot 1 includes a first arm 11. which is pivotally supported on a base 2 and rotates in a horizontal plane.
A second arm 12. which is pivotally supported by the tip of the first arm tt and rotates in a horizontal plane. A rotary direct-acting link 13 that is pivotally supported at the tip of the second arm 12 and rotates once and moves in the axial direction. and a hand (hand effector) 14 fixed to a rotary translation link 13. At the tip of the second arm 12 is a two-dimensional visual recognition camera (
Visual device) 7 is installed. The output video signal of the camera 7 is sent to an image processing device 8, where it is processed and the characteristics mW (for example, center coordinates) of the workpiece etc. are calculated. The image processing device 8 is connected to a robot controller 9. The robot controller 9 is connected to the arm 11 of the robot 1. 12. Controls the driving of the link 13 and the like. Arm 11.12. Link! The third motor is driven by a known servo motor. The robot controller 9 and the image processing device 8 are composed of a CPU, memory, interface, etc.

FIG. 2 is a diagram of the joint structure of the robot arm in the robot 1 described above, in which the first axis Kl, the second axis 2 and the fourth axis 4 are rotary joints, and the third axis is a prismatic joint 3. .

The robot 1 is given in advance data regarding the position and orientation of the 1-1 gauge point at which the robot 1 should grip the workpiece Wk with its hand 14. However, the work Wk is 1
, there is no guarantee that the position and orientation of the workpiece W will be properly positioned. It is necessary to fix the points.

Second training! (2)--The operation of the robot 1 from the time it starts until it grips the workpiece Wk with the hand 14 will be described next. FIG. 6 shows an outline of the processing procedure performed by the controller 9 for controlling the operation of the robot 1.

Referring to FIG. 6, an operation is started for driving arm II, 12 to move hand I4 toward a predetermined reference point II (step 21). This is a PTP (point-to-point) motion for 11 joint angle joint angles that has been commonly performed in the past. Then, while moving, the workpiece Wk is placed within the field of view of camera 7.
is constantly checked to see if it has been entered (step 22).
). When the workpiece Wk enters the field of view of the camera 7, the control system is switched to a camera coordinate control system as described in detail below (step 23).

In FIG. 4, the field of view 73 of the camera 7 is shown.

The solid line image indicated by the symbol WI11 is the work image,
The center position (x, y) is detected by the image processing device 8 based on the video signal from the camera 7. The position of the workpiece image Wm is the arm 11. 12, that is, the movement of the camera 7, changes within the field of view. That is, the value of its central coordinates (x, y) changes.

The center position of the image of the broken line indicated by the symbol Wo (X d.

yd) represents the desired workpiece position obtained from the camera coordinates, and this becomes the target position.

Further, this center position (x, y) is also the center position of the hand 14 within the camera coordinates d. Center position (
x, y) are values known in advance from the geometric relationship between the camera 7 and the hand 14, and are given to the robot controller 9.

Work Wk is camera 7's l! After entering the Pf7S, the control system executed by the controller 9 is as shown in FIG. FIG. 3 is a block diagram showing the processing executed by the controller 9. As shown in FIG.

The center position (x, y) of the workpiece Wk is given from the image processing device 8 as the foot information as described in Section 1 above. 1-1 target values (x, y,) are set in advance.

First, the following calculation is performed.

ΔX"X, -X (1) ΔY""'W dY Further calculations are performed using ΔX and Δy obtained by this calculation.

Here, J-1 is a Jacobian matrix that establishes the relationship between displacement in the camera coordinate system and displacement in joint coordinates, and can be easily determined using joint angles. The joint angle is determined by the encoder El, which detects the position of the arm 11.12 (joint Kl, 2).
Pj by E2.

θ, 0 obtained in this way is the workpiece image Will
The center position (x, y) of is the camera constellation + term (X d.

The first barrier flK to approach yd)! , 2 at the second joint
<Arm 11, 12) has a meaning as a speed command value.

For these θ and θ, encoders E1 and E2
The velocity information obtained by differentiating the position information obtained from (S is a Laplace operator and means differentiation) is fed back, and the deviation is used as a command value to the servo amplifiers AI and A2. Servo amplifiers AI and A2 are arm II and +2, respectively.
A current according to the command value is supplied to the motors Ml and M2 that drive the motors Ml and M2.

As described above, the above control is performed every time an image is captured by the camera 7, and the center position of the workpiece Wk is set to (x, yd
), the arms 11 and 12 are controlled so that

” In the explanation of the two legs, the center position (x.

Although y) is used as feedback information, it is also possible to calculate equation (1) in the image processing device 8 and set ΔX and Δy as the 1-11 values of the control system. In either case, the functional block diagram of the control system is shown in FIG.

The above is the control regarding the position, but first, the attitude of the work Wk will be described. Center position of hand 14 and workpiece W
Even if the center positions of the work Wk and the hand 14 do not match, accurate gripping cannot be performed unless the posture of the workpiece Wk and the posture of the hand 14 match.

To match the hand postures, the rotary translation link 13 (m
This is done by rotating 4) at the 4 joints.

That is, the workpiece posture with respect to the base coordinate system fixed to the base 2 of the robot 1 is measured, and the fourth joint 4 is rotated to match the posture.

The workpiece posture can be determined precisely as follows. Referring to FIG. 5, the camera coordinates X of the workpiece image Wffl. Yo
The image processing device 8 determines the inclination angle α with respect to the angle α. Next, the arm It obtained from encoders El and E2,
12 joint angles θ and θ2, calculate the camera coordinates XY and the C inclination angle β of the base coordinates
The angle (that is, the posture) with respect to the base coordinates X8Y8 is determined.

Returning to Figure 6, what are the target values (x, y)? If the II constant values d (x, y) match, the workpiece will be located at the true position of the hand 14, so the link 13 (third function 1ii
jK3) is lowered (step 25). At this time, if necessary, the posture change control of the hand 14 described in -1- is performed. Then, by closing the hand I4, the workpiece Wk is gripped.

After the workpiece Wk enters the field of view of the camera 7 as described above, by switching to a control system based on the camera coordinate system, calculation processing can be performed by performing equations (1) and (2). Since complicated coordinate transformation calculations as in the conventional method are not required, high-speed processing and high-speed operation are possible. Positioning accuracy is also improved.

In the above embodiments, a planar work type robot configured with rotary joints has been described, but other types of back surface work type robots such as a Cartesian coordinate type and two cylindrical coordinate type are also applicable.

Furthermore, the present invention can be similarly applied to three-dimensional working robots.

[Brief explanation of drawings]

Figure 1 is a perspective view showing the entire robot system, Figure 2 is a perspective view showing the entire robot system.
The figure is a diagram of the arm joint structure of the robot in Figure 1, Figure 3 is a block diagram depicting control processing from a functional perspective, Figure 4 is a diagram showing the coordinate system of the camera and the image captured in its field of view.
Fig. 5 is a diagram showing the relationship between the camera coordinate system and the base coordinate system, and Fig. 6 is a flowchart showing the processing procedure by the controller.
It is a chart. 1...Robot, 2...Base. 7...Camera, 78...Camera field of view. 8... Image processing device. 9...Robot controller. ■, 12...Arm, +3...Link. 14...hand. Kl, 2. 3. 4...Joints. Wk...work, Wa+...work image. Wo...Target position. Applicant for the above patent: Tateishi Electric Co., Ltd. Agent: Kenji Ushiku (and one other person) Figure 1, Figure 2, 2

Claims (1)

[Scope of Claims] A visual device provided to maintain a predetermined positional relationship with a hand effector held on an arm member of a robot, and a workpiece image position grasped by the visual device and the predetermined position. a first calculation means for calculating, in the coordinate system of the visual device, a quantity related to the deviation from the target position determined by the positional relationship; and means for driving and controlling the arm member so that the position of the hand effector approaches the target position according to the predetermined amount determined by the second calculation means. Control device.