JPS6190205A - Compensating method of absolute positioning error of robot - Google Patents

Abstract

PURPOSE:To improve the accuracy of positioning of absolute position of robot, by correcting the positioning error against the absolute coordinates of a robot by detecting the error from a correction table in which each position of robot is specified. CONSTITUTION:The motion range of a robot is divided into gridiron, and a correction table is prepared to each division specifying correction volume of DELTAA-DELTAC, or correction volume for rectangular coordinates system deltax-deltaz. Suppose that the recognition of the present position of robot is P0(x',y',z'), against which the robot is actually positioned at P1(x,y,z) having differences of deltay, deltaz. Denoting that the rotating angle of rotation axis 8 obtained from encoder is A, and angles of 4th and 5th joints are B' and C' respectively, B' and C' should be corrected by DELTAB and DELTAC to make a correct recognition of present position, and B=B'+DELTAB; C=C'+DELTAC are obtained. When (x), (y), (z) coordinates are decided according to these angles A-C, then; x=f(A,B,C); y=g(A,B,C); and z=h(A,B,C), and coincide to the actual position P1(x,y,z).

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a method for compensating absolute positioning errors of a robot.

[Conventional technology]

In general, industrial robots memorize the work order, position, and other information by having the operator teach the robot by moving the robot arm in advance, and then repeat the work by replaying the memorized information. The playpa, ta method is adopted. In the case of this method, the mouth?
Accurate movement function (absolute precision) in the absolute coordinate system of the cut is not required, but it is
The above absolute precision is essential for a robot that performs work in accordance with the output (computerized design).

Errors in the absolute coordinate system are less likely to occur in the case of orthogonal mouths, but multi-jointed mouths? There are some cuts, especially vertical ones.

(1) The actual arm length (designed arm length) is different (manufacturing error).

(2) Preset mouth? The state of the robot's coordinate origin differs from the state of the actual robot's coordinate origin (assembly/adjustment error).

(3) Differences between the computational geometric model and the actual robot (
twist of the rotating shaft, manufacturing errors).

(4) Reducer based on the weight of the arm and the weight of the workpiece,
Uneven rotation of the rotating shaft due to deformation of the bearing, deflection of the arm, manufacturing errors, or unavoidable structural causes.

[Problem that the invention seeks to solve]

That is, even if the logo receives position data in an absolute coordinate system, it cannot be accurately positioned at the position indicated by the position data within the entire operating space range.

The present invention has been made in view of the above-mentioned circumstances, and an object of the present invention is to provide a method for compensating absolute positioning errors of a robot, which improves positioning accuracy with respect to absolute coordinates.

[Means for solving the problem] [According to the present invention, the operating range of the robot is divided into appropriate intervals, and for each division, the positioning error with respect to the absolute coordinates or this error is converted into the amount of movement of each axis of the robot. A memory table is prepared that stores the value as a correction value, and when there is a movement command based on the position of the robot in the absolute coordinate system, the correction value stored in the correction table corresponding to that position is prepared. is used to compensate for robot movement position errors.

[For production]

The positioning error with respect to the absolute coordinates of the opening cut is read out from the correction table and compensated for.
Even for robots with large errors in the absolute coordinate system, positioning accuracy with respect to the absolute coordinates can be improved.

〔Example〕

Hereinafter, the present invention will be described in detail with reference to the accompanying drawings.

FIG. 1 is a perspective view showing an example of an industrial opening cut. This robot has a first arm 11, a second arm 2, and a wrist 3.
This vertically articulated robot has a vertical articulated robot.
It is constituted by a combination of unit mechanisms: a joint 4 and a second joint 5, a third joint 6 and a fourth joint 7 that twist and bend the wrist 3, and a pivot shaft 8 that rotates the entire robot.

We will explain the case of compensating for errors caused by the weight of the arm and workpiece when positioning with respect to absolute coordinates in step &2) above. Figure 2 is a schematic diagram of Figure 1, and the weight of the arm and workpiece. A moment is applied to the bow, arm, and joint, and due to this moment, the first joint 4 rotates by ΔB, and the second joint 5 rotates by ΔB.
is rotated by ΔC, and the position determined by the robot has an error of δa and δ2 compared to the ideal state (arm and joint stiffness is infinite (see dashed line)). That is, the absolute positioning error at this position is δa.

It becomes δ2.

Therefore, in order to compensate for the above errors δa and δ2t, the deformations ΔB and ΔC due to the moment are determined in advance by measurement, etc., and then the deformations ΔB and ΔC due to the moment are determined by measurement etc. The obtained angles B and C may be corrected.

Mouth now? Recognition of the position determined by the cut is performed using P@(
x', y', z'), an error of δa, δ2 occurs with respect to the actually positioned position P1 (x, y, z).

If the correct recognition degrees of this current position P□(χe7eZ> are B' and C', respectively, B/ and C' are ΔB,
Corrected by ΔC, BmiB'+ΔB CmiC'+ΔC Calculate the xyz coordinate position based on angle A and corrected angles B and C. *C) 2mih(B, () Matches the actually positioned position P1 (X#7#Z).

On the contrary, positioning to position P10C*71 specified by xyz coordinates is expected to deform the ΔB and ΔC joints,
Positioning is done by P@G" + 7' as shown in Figure 3.
yz / ). That is, the position P□(X p
Find each angle A, B, and C from 7 p z), A== f'< ) Among these angles, B and C are corrected by ΔB and ΔC, and each joint of the robot is controlled so that it becomes A, B', and C'. , the ΔB and ΔC joints are deformed to B and C, and the position P□ (Xsy
, z).

Also, the above ΔB and ΔC vary depending on the moment applied to each joint of the arm, but in the case of a vertically articulated robot, the mouth? Since the moment changes depending on the posture of the kit, ΔB
, ΔC are not constant, but change depending on the magnitude of Bpc. Therefore, we divided B and C appropriately within the operating range, and
As a two-dimensional matrix of C, a correction table shown in the following table is created.

Note that in the case of the rotation angle A of the rotation axis 80, since it does not depend on the postures of the first arm 1 and the second arm 2, the correction amount ΔA is one-dimensional due to A. When the correction amount of a certain axis is affected by the position of another axis, the correction table becomes a two-dimensional or three-dimensional matrix depending on the axis that is affected.

In addition, as shown in FIG. 4, the operation area of the opening and opening is divided into basic grid shapes, and ΔA and ΔB are determined for each division.

Compensators that each store ΔC (the turning direction correction amount for four people) or errors in the orthogonal coordinate system (δ-work, δ-A, δ-work)
A regular table may also be created.

In addition to the correction table that stores correction values for each appropriately divided area of the operating range as shown in Table 1, the correction table stores correction values for each appropriately divided area of the operating range. I can think of things that I remember. In the latter case, the correction value for positions other than the above divided positions is
IJ can be determined by a rough interpolation method based on the correction values at the divided positions before and after that position.

Furthermore, the method of the present invention is applicable not only to vertically articulated robots but also to other types of mouths. It can also be applied to

〔Effect of the invention〕

As explained above, according to the present invention, the positioning error with respect to the absolute coordinates of the location and the location is read out from the correction table stored corresponding to each position and the error is corrected. It is possible to improve positioning accuracy.

[Brief Description of the Drawings] Fig. 1 is a perspective view showing an example of an industrial gutter, and Figs. 2 and 3 are the grate of Fig. 1 used to explain the method of the present invention. FIG. 4 is a diagram showing an example of the motion range of the robot shown in FIG. 1 in a vertical plane and the classification of the motion range when creating a correction table according to the present invention. 1... 1st arm, 2... 2nd arm, 3... wrist, 4... 1st joint, 5... 2nd joint, 6... 3' joint, Fig. 1 Figure 2 Figure 3 2 Figure 4

Claims (1)

[Claims] The robot's operating range is divided into appropriate intervals, and the error between the position taught in the robot's absolute coordinate system within each division and the position when the robot is actually positioned based on that position, or this error, is calculated for each axis of the robot. A correction table is prepared in which the value converted to the amount of movement is stored as a correction value for each category, and when there is a movement command based on the position in the absolute coordinate system of the robot, the correction is performed in accordance with the position. A method for compensating absolute positioning errors of a robot, characterized in that errors in the movement position of the robot are compensated for using correction values stored in a table.