JPH08174453A - Measuring device of positioning error in robot device and correcting method for positioning error - Google Patents

Abstract

PURPOSE: To easily find a formula for showing the positioning error, taking in account a link parameter of an arm of a robot device, so as to easily estimate the positioning error by finding respective influence coefficients and respective link parameter errors on the basis of an action command value and the distance measured by a reference body. CONSTITUTION: A coordinate of the action target position of a robot 1 is input into a computer device 19, a positioning coordinate of a X-Y table 13 is calculated, and the calculated coordinate is sent to as a command value to a table controller 14. The X-Y table 13 is moved according to the command value, and a Z-target 17 above the table is moved to the specified position. A measuring body 18 provided on the tip of the arm 3 of the robot 1 is approached to the target 17 according to an NC data 15, the joint angle is calculated by the computer 19, and the influence coefficient wherein respective link parameters affect the positioning error is calculated. After the action of the robot 1 is completed, the positioning error dN is measured by the measuring body 18.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a positioning error measuring device and a positioning error correcting method in a robot device.

[0002]

2. Description of the Related Art Conventionally, when a positioning error of a robot before shipment or a robot used at a work site is corrected, an actual robot is operated and the deviation amount of positioning at that time is calculated. It was done by the operator's understanding of the experience. That is, based on the experience of the operator, by changing the parameters inside the control program or by modifying the actual drive mechanism,
The performance of the robot was maintained and improved.

[0003]

According to the method of correcting the positioning error of the robot as described above, the operator determines the positioning error of the robot for each robot and, based on this determination, the control program. There has been a problem that a lot of experience is required to change the internal parameters and the work of correcting the drive mechanism of the robot, and a lot of time is required for the correction work.

Therefore, it is an object of the present invention to provide a positioning error measuring device and a positioning error correcting method in a robot apparatus which can solve the above problems.

[0005]

In order to solve the above-mentioned problems, a positioning error measuring apparatus in a robot apparatus according to the present invention is a positioning error measuring apparatus in an articulated robot apparatus having an arm composed of a plurality of links. ,
A reference body having a reference plane in the XYZ directions is arranged on one of the XY table or the arm tip of the robot apparatus, and a measuring body for measuring the distance to each of these reference planes is arranged on the other side, and the robot apparatus is provided with A control device that outputs a motion command and a plurality of motion error values and the distance measured by the reference body from the control device, and the positioning error of the robot arm tip is calculated by adding the respective influence coefficients. And an arithmetic processing unit that obtains each influence coefficient and each link parameter error by using an expression represented by the sum of the link parameter errors.

Further, the method of correcting the positioning error in the robot apparatus of the present invention estimates the positioning error of the arm tip of the robot apparatus based on the sum of products of the influence coefficient and each link parameter error obtained by the above measuring apparatus. In this method, each link parameter error that eliminates the estimated value is obtained, and the arm operation command value of the robot apparatus is corrected based on the obtained link parameter error.

[0007]

According to the above-mentioned positioning error measuring device, an expression representing the positioning error in consideration of the link parameter of the arm of the robot device can be easily obtained for each robot device. Can be easily predicted.

Further, according to the above-described method for correcting the positioning error, the positioning error of the arm tip can be predicted based on the influence coefficient and the link parameter error obtained by the measuring device for the positioning error, so that the predicted positioning error disappears. As described above, it is sufficient to correct the operation command value of the arm, and the correction can be performed very easily.

[0009]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to FIGS. First, the principle of measuring the mechanism error in the articulated robot apparatus will be described based on the link model shown in FIG.

In order to identify a mechanical error of a robot apparatus (hereinafter, simply referred to as a robot) 1 having an arm 3 composed of a plurality of links 2, it is necessary to model the robot 1, but the mechanical error can be best expressed. There are those shown in the following formulas (1) and (2).

D _N = M ₁ Δθ + M ₂ Δb + M ₃ Δa + M ₄ Δα + M ₅ Δβ ... (1) δ _N = M ₂ Δθ ++ M ₃ Δα + M ₅ Δβ ... (2) As shown in FIG. 1, five link parameters (α _i , β _i , θ _i , a _i , b _i ) are adopted so as to easily correspond to the mechanism.

Where α _{i is the} inclination of the rotating shaft with respect to the mounting surface β _{i is the} inclination of the rotating shaft with respect to the mounting surface θ _{i is} the rotation angle of the rotating shaft around the axis a _{i is the} inter-axis distance b _{i is} the length in the rotating shaft direction It is.

As shown in the above equations (1) and (2), the error in the tip position coordinate of the arm 3 of the robot 1 is the sum of the influences of the errors of each link 2. In each of the above formulas,
When the number of joints of the robot is N, M _i is a coefficient of a 3 × N matrix (hereinafter, referred to as an influence coefficient) indicating how much the parameter error of each link 2 influences the shift of the Nth link coordinate, These are uniquely determined by the relationship between each link parameter and the Nth link.

Further, Δθ, Δα, Δβ, and Δb represent the parameter error of each link parameter and are represented by an N × 1 matrix. In addition, d _N is a minute translation (d
_{x, d y, d z)} , δ N represents microspheroidal ([delta] _x of the N links coordinate, [delta] _y, the [delta] _z).

Here, a method of _obtaining the influence coefficient M _i will be described. As shown in FIG. 1, (i-
1) Considering a 4 × 4 transformation matrix A _i for transforming from the coordinate system to the i coordinate system, the five parameters existing in the link are slightly changed (Δθ _i , Δα _i ,
The small change dA _i of the conversion matrix when Δβ _i , Δa _i , Δb _i ) is expressed by the following equation (3) by linear approximation.

[0016]

[Equation 1]

Further, the minute change in the i coordinate system is expressed by the following equation (4) using the minute translation / rotation conversion matrix Δ. dA _i = A Δi (4) where Δi is a minute translation vector * d _i (d _x , d
_y , d _z ) (* indicates that the character following it is a vector, and the same applies hereinafter), and a minute rotation vector as * δ _i
Letting (δ _x , δ _y , δ _z ) be expressed as in the following equation (5).

[0018]

[Equation 2]

Here, in the case of an articulated robot as shown in FIG. 2, a homogeneous transformation matrix (T) for transforming from the robot reference coordinate system {0} to the position and orientation of the robot arm tip coordinate system {6}. Considering that the minute change in ₆ ) is caused by only a certain coordinate system {i}, dT ₆
Is expressed by the following equation (10).

[0020]

(Equation 3)

Here, * n, * s, * a, and * p are homogeneous transformations from the coordinates of the tip of a certain (i-1) link to the coordinates of the tip of the robot, and are given by the following equation (15). Represented by ⁱ T ₆ = A _i A _{i + 1} A _{i + 2} ... A ₆ (15) Further, ^i-1 T ₆ is expressed by the following equation (16).

[0022]

[Equation 4]

Therefore, the vectors of the microtranslation ^T6 * d and the microrotation ^T6 * δ due to the microchanges of all coordinate systems of the robot are the sum of the microtranslation and microrotation vectors caused by the microchanges of the respective coordinate systems. Because there is
They are expressed by the following equations (17) and (18), respectively.

[0024]

(Equation 5)

Therefore, the above (8), (9) and (17), (18)
According to the equation, each parameter of Δθ, Δα, Δβ, Δa, and Δb exists by the number of joints of the robot arm, and the mechanism error corresponding to each of the joints is calculated in the XYZ direction using the influence coefficient. It is expressed by the following formulas (19) and (20).

[0026]

(Equation 6)

Note that M _j ⁱ = [M _j ¹ , M _j ² ... M _j ⁶ ]
Therefore, in the case of a 6-axis robot, M _j ⁱ becomes a matrix of 3 rows and 6 columns, and Δθ, Δα, etc. are represented as follows.

Δθ = [Δθ ₁ , Δθ ₂ ... Δθ ₆ ] In this way, the influence coefficient and the link parameter error can be obtained. Therefore, it can be seen from the above equations (1) and (2) that the positioning accuracy of the tip of the arm 2 of the robot 1 is a dependent variable of each link parameter error.

From this, the unknown link parameter error can be obtained by measuring the positioning accuracy of independent points that can determine each of these unknown link parameters.

Thus, if the influence coefficient and the link parameter error are known, the positioning error of the tip of the arm 3 can be known. Next, using the above principle, an error measuring device in the robot device will be described with reference to FIGS. 3 and 4.

FIG. 3 is a diagram showing the overall structure of a measuring device for measuring the mechanical error in the robot device, and FIG. 4 is a perspective view showing the structure of the main parts of the measuring device. The measuring device 11 includes a mounting table 1 on which a robot (test robot) 1 is mounted on one end side.
2, an XY table 13 mounted on the other end of the mounting table 12, a table controller 14 for controlling the position of the XY table 13, and the robot 1 based on the NC data 15 for testing. As shown in FIG. 4, the control device 16 that controls the X-Y table 13 is disposed on the X-Y table 13 side.
A target (also referred to as a cubic) 17 having three reference planes 17a in the YZ direction and three reference planes 17 of the target 17 as an end effector of the robot 1.
a shape having a measuring surface 18a corresponding to a (female mold 3
A sensor (not shown, for example, a laser sensor or the like is used) for measuring the distance to each reference surface 17a, and the measurement object 18 attached to each measurement surface 18a; Of the detection NC signal and the test NC data 15 input to the control device 16, and performs a predetermined calculation, that is, a calculation based on the above formulas (1) and (2). Has been done.

Next, a method of measuring the mechanical error of the robot using the above measuring device will be described. Briefly explaining, the target 17 is moved to a predetermined position by the XY table 13, the measuring body 18 on the side of the robot 1 is moved to a position corresponding to the target 17, and the measuring body 18 is moved. With the 6 sensors attached,
This is a method in which the distance between each reference surface 17a of the target 17 and the measuring body 18 is measured to obtain a positioning error.

A detailed description will be given below. First, the positioning coordinates of the XY table 13 are calculated by inputting the coordinates of the operation target position of the robot 1 to the computer device 19, and the calculated coordinates are sent to the table controller 14 as command values.

According to this command value, the XY table 1
3 is moved, and the target 17 placed on this is moved to a predetermined position. Next, the robot 1 causes the measuring object 18 attached to the tip of the arm 3 to approach the target 17 according to the NC data 15. At the same time, on the computer 19 side, the joint angle when the robot 1 approaches the target 17 is calculated according to a predetermined motion, and each link parameter error affects the positioning error according to the equations (1) and (2). The influence coefficients (matrix) M _{1 to} M ₅ are calculated.

After the operation of the robot 1 is completed, the measuring body 18 measures the positioning error d _N. After the measurement is completed, the target 17 and the robot 1 are returned to the predetermined positions. By repeating such a procedure, the positioning error at each position is measured, and the measurement results d _N at that time, the respective link parameter errors, and the influence coefficients M _{1 to} M ₅ indicating the correlation with the positioning error are stored in a file. save.

Next, the positioning error of the arm 3 of the robot 1 is corrected based on the influence coefficients M _{1 to} M ₅ and the link parameter errors calculated as described above.
That is, as shown in the block diagram of FIG.
The joint angle of the arm 3 is calculated, and based on this joint angle, the influence coefficients M _{1 to} M ₅ and each link parameter error, the positioning error due to the NC data which is the operation command of the arm 3 becomes (1) and (2). Calculated from the formula. Therefore, based on the obtained positioning error, the correction is performed so as to obtain the original commanded position.

Then, the joint angle of the arm 3 of the robot 1 is calculated based on this corrected command value, and this joint angle is commanded to the robot 1. Thus, the positioning error can be easily eliminated.

According to this correction method, the parameter error is indirectly reflected as the correction amount, so that it is possible to improve the accuracy of the positioning error due to the mechanical error without solving a complicated inverse conversion formula. .

Generally, the robot commands the position and orientation of the robot's arm tip in a Cartesian coordinate system, and converts it into the actuator coordinate system of each axis of the robot by the inverse transformation formula, whereby the joint angle of each axis of the robot is converted. Are in control.

By the way, normally, in the inverse conversion equation for obtaining the joint angle θ, the link lengths a and b shown in FIG. 1 are treated as arithmetic constants. On the other hand, the link mounting angles α and β are originally
Although it exists in the form of a trigonometric function in the formula, its design value is generally 90 degrees or 0 degrees, so it is omitted in the formula. As a result, the inverse transformation formula becomes simple and the solution can be easily obtained. However, when trying to find the solution of the inverse transformation equation considering the mechanical error, the mounting angles α, β
Also has to be treated as an arithmetic constant, which makes the inverse transformation equation very complicated and has the drawback that the solution cannot be easily obtained. The above correction method can eliminate such a defect.

Next, an actual example of correction will be described. FIG. 6 shows the result of actually measuring the positioning accuracy of the arm end of the robot by the above measuring device. As shown in FIG. 6, the measurement points were measured at a total of 36 points with the posture of the measuring body 18 kept constant. The measurement result of FIG. 6 shows the positioning error decomposed for each direction of the robot reference coordinates.

In this measurement, the payload is 3
A 6-axis robot with kg, upper arm 240 mm, and lower arm 250 mm was used. From this measurement result, the absolute positioning error of this robot was measured to be 4.6 to 8.1 mm.

[Table 1] shows the identification result of each parameter error of the robot calculated from the measurement data shown in FIG.

[0044]

[Table 1]

The result of applying the above-described correction method will be described. That is, the positioning error was calculated from the identification result of the mechanical error shown in [Table 1], and NC data corrected for the error was given to the robot to measure the actual positioning accuracy. This result is shown in FIG. FIG. 7 shows the measurement result of the positioning accuracy when the above-described correction method is applied to the measurement point of Z = −3.0 mm in FIG.

From this result, the positioning error when the correction command value was given was measured to be 2.1 to 4.3 mm, and it was confirmed that the accuracy was improved by about 50% as compared with that before the correction. In the above description, the target is positioned by using the XY table. However, if the target is attached to the probe section of the coordinate measuring machine, it is possible to achieve a higher degree of accuracy in the robot. It is possible to measure various positioning errors.

[0047]

As described above, according to the positioning error measuring apparatus of the present invention, the expression representing the positioning error in consideration of the link parameter of the arm of the robot apparatus can be easily obtained for each robot apparatus. The positioning error of the arm end of the device can be easily predicted.

According to the positioning error correction method of the present invention, the positioning error of the arm tip can be predicted based on the correction coefficient and the link parameter error obtained by the positioning error measuring device. It is only necessary to correct the arm operation command value so as to eliminate the positioning error. Therefore, it is not necessary to convert the positioning error into the correction amount in the arm drive coordinate system, and correction can be performed very easily. You can

[Brief description of drawings]

FIG. 1 is a perspective view showing a general link model for obtaining a positioning error in a robot apparatus according to an embodiment of the present invention.

FIG. 2 is a perspective view showing a specific link model for obtaining a positioning error in the robot apparatus of the same embodiment.

FIG. 3 is an overall view showing a schematic configuration of a positioning error measuring device in the robot apparatus of the embodiment.

FIG. 4 is a perspective view of a main part of the measuring device according to the embodiment.

FIG. 5 is a block diagram showing a correction procedure in the embodiment.

FIG. 6 is a diagram showing a measurement result of a positioning error of the robot apparatus by the measuring apparatus according to the embodiment.

FIG. 7 is a diagram showing a result when the measurement result of FIG. 6 is corrected.

[Explanation of symbols]

 1 Robot Device 2 Link 3 Arm 11 Measuring Device 13 XY Table 15 NC Data 16 Control Device 17 Target 17a Reference Surface 18 Measuring Object 19 Computer Device

Claims (2)

[Claims] 1. A measuring device for a positioning error in an articulated robot apparatus having an arm composed of a plurality of links, wherein a reference having a reference plane in the XYZ directions is provided on an XY table or one of the arm tips of the robot apparatus. A control device that places a body on the other side and a measurement body that measures the distance to each of these reference planes, and outputs an operation command to the robot device, and measures with this operation command value and the reference body from this control device. By inputting the calculated distance, the positioning error of the robot arm tip is expressed by the sum of a plurality of link parameter errors, each of which is calculated by the effect coefficient. And a processing device for calculating an error in each link parameter, and a device for measuring a positioning error in a robot device. 2. A link for estimating the positioning error of the arm tip of the robot device based on the product sum of the influence coefficient and the error of each link parameter obtained by the measuring device according to claim 1, and eliminating each estimated value. A method for correcting a positioning error in a robot apparatus, which comprises obtaining a parameter error and correcting an arm operation command value of the robot apparatus based on each obtained link parameter error.