JPH012104A - Robot positioning error correction method - 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] [Field of Industrial Application] The present invention relates to a method for correcting positioning errors of a robot in a method in which the robot is taught offline using a CAD/CAM system to control the position and orientation of the robot; In particular, it relates to improving absolute positioning accuracy.

[Prior Art] Industrial robots are used for welding, assembly, painting, etc. In order to fully automate this robot and prevent collisions with design components, robot and software off-line teaching, which simulates work using a CAD/CAM system and teaches an actual robot, is attracting attention.

In this case, absolute positioning accuracy is required as the positioning accuracy of the robot, not position repeatability. Absolute positioning accuracy is an accuracy that represents the amount of deviation of the position and posture reached by the actual robot with respect to the commanded position, posture, or angle of each axis of the tip end of the robot. The absolute positioning accuracy of this robot is not guaranteed, and when performing offline teaching using a CAD/CAM system,
In a CAD/CAM system, the position/orientation or angle command values for each axis are created assuming that there is no absolute positioning error of the robot, so errors in hand position and orientation occur.

Robot hand position in CAD/CAM system.

The cause of the error in posture is the geometric difference between the teaching data on CAD/CAM and the actual motion data of the robot. That is, as shown in FIG. 7, the robot 71 on the screen displayed on the CAD/CAM graphic display has the joint origin of θ-0 and each link length as the design values of the robot 72 shown in the design drawing. L1. L2 is defined with high precision, and the robot's movement data is also CAD/
Accurate values are generated on the CAM. However, when manufacturing the actual robot 73, manufacturing errors occur, resulting in an origin error Δθ of each axis position detector and each link length error ΔL, ΔL2.
In addition, base errors Δ100 sometimes occur during installation, and these errors cause geometric differences between the graphic model on the CAD/CAM and the actual model, resulting in errors in the position and posture of the robot's hands.

Several proposals have been made to correct this error. For example, Journal of the Robotics Society of Japan, Volume 3, No. 2 (1986)
rN described on pages 64 to 69 of
AcI (I CATS Offline Teaching System) or the paper presented at the 48th Technical Information Processing (TIPS) Study Group held on July 3, 1986, "Offline Programming System for Spot Welding Robots"
A method for correcting the positioning accuracy of a robot hand by measuring the position and orientation of the robot hand as disclosed in (hereinafter referred to as conventional example 1).

Or JP-A-82-28808, JP-A-62-
A method of correcting the positioning accuracy of a hand by measuring the distance between a plurality of calibration positions and each joint angle at each calibration position, as disclosed in "Robot Coordinate System Calibration Method" described in Publication No. 28809. (hereinafter referred to as conventional example 2),
Alternatively, as disclosed in ``Method for correcting the origin of an articulated robot'' published in Japanese Patent Application Laid-open No. 59-205282, positioning is performed in two ways at the same point using a symmetrical arm, and the positioning is performed in two ways. A method (hereinafter referred to as conventional example 3) has been proposed in which the deviation is detected and the origin is corrected to improve accuracy.

In the method disclosed in Conventional Example 1, a matrix [RF] indicating position/orientation data actually operated by a real robot is created for a matrix [WF] indicating position/orientation teaching data generated on CAD/CAM. The transformation matrix between the matrices [WF] and [:RF] representing these two data is determined as the position/orientation transformation matrix [RW] using the following formula, and the transformation matrix is calculated on the CAD/CAN for each teaching point. The positioning accuracy of the robot is corrected by multiplying the teaching data Nico's matrix [RW]. This matrix [RW] is the posture of the robot,
Since each teaching point is different, interpolation is performed to improve accuracy.

In addition, the method disclosed in Conventional Example 2 creates simultaneous equations using the measured distance between the calibration positions, the joint angle at each calibration position, and the link length at the time of design, and solves this equation to calculate the link length. The positioning accuracy of the robot is corrected by determining the error Δ°C and the installation error Δj of each joint.

[Problems to be Solved by the Invention] In the method of Conventional Example 1 above, the influence of the link length error Δπ and each joint attachment error Δb on the robot hand position and posture is determined for each robot posture, that is, for each teaching point. Since the robot hand positioning error is different, it is not possible to effectively correct the robot hand positioning error unless a position/orientation transformation matrix is obtained for each teaching point, and if accurate correction is attempted, the amount of data required becomes enormous. There was a point.

Further, even if interpolation is performed to improve accuracy, only approximate values can be calculated, and accurate correction cannot be performed.

In Conventional Example 2, the angle of each joint at each calibration point of the robot is measured by a joint angle detector inside the robot body, so in general general purpose robots where the internal state of the robot is not disclosed, each joint angle There was a problem that this method could not be applied because it could not be measured.

In addition, in the method of Conventional Example 2, the coordinate system on the base of the robot body, that is, the robot coordinate system, the coordinate system of the coordinate measuring machine that measures the robot hand position, that is, the measuring machine coordinate system, and the actual work The robot has three types of coordinate systems: the work coordinate system, which is the coordinate system of the object to be measured, and the correction of the robot's hand position is performed on the measuring machine coordinate system. Since the values of the corrections for the robot's hand position differ depending on how the reference coordinates are taken, corrections must be calculated every time the measuring machine coordinate system and the work coordinate system do not match. There are also problems.

In the method of Conventional Example 3, the positioning deviation is detected by the position data of each joint, that is, the encoder value, or the encoder value of each joint of the industrial robot is not disclosed to the public, and the robot model The problem is that it cannot be applied to general-purpose robots because it is limited to SCARA type robots.

Furthermore, in addition to measuring the position of the robot's hand, there are also methods to measure its posture, such as by using multiple cameras or by equipping the robot's hand with three light emitting diodes and determining the inclination of the hand from the photographed image. A complicated system is required for measurement, and there is still a problem with versatility.

This invention was made to solve this problem, and it is possible to improve the number of robots that operate based on the position/orientation or angle command values of each axis from a CAD/CAM system, etc., without using a robot internal sensor. The purpose of this invention is to propose a robot positioning error correction method that can correct absolute positioning errors based only on relative position information of points.

[Means for Solving the Problems] The robot positioning error correction method according to the present invention includes: (a) command values for a plurality of positions/postures or angles of each axis (
Hereinafter, these are referred to as position/orientation command values. ) to the robot control device, (b) actually measure the position of the robot's hand in the robot's base coordinate system corresponding to the command values for each of the above positions and orientations, and (c) give the command values for each of the above positions and orientations. Then, calculate the origin error of each axis of the robot using a micro-relationship equation from the relative value of the actual measured value of the hand position, (d) Correct the command value of the 0 position and orientation with the origin error of each axis mentioned above, It is characterized by controlling the position and posture of the robot.

[Operation] In this invention, the origin error of each axis of the robot, which has a large influence on errors, is determined by the command values of multiple positions and orientations, and the measured value in the robot coordinate system of the relative position at which the robot's hand position moves based on the command values. By determining this, we aim to improve the absolute positioning accuracy of the robot.

[Embodiment] FIG. 1 shows a block diagram of an apparatus according to an embodiment of the present invention, in which 1 is a CAD/C that performs offline teaching.
AM system, 2 is a graphic display of the CAD/CAM system, 3 is a post-processor for implementing the method which is the main part of the present invention, 31 is a storage means, 3
2 is a micro-relationship calculation means, 33 is an arithmetic means, 4 is a robot, 41 is a control device of the robot, 42 is a mechanical part of the robot, and 5 is an input device for inputting data to a post processor, which is actually measured. This is used to input the hand position data of the robot.

Hereinafter, before explaining the operation of this embodiment, the position of the hand of the robot 4 will be measured using a three-dimensional measuring device, and the deviation amount of the rotation origin of the robot joint, that is, the origin error Δj of each axis position detector will be calculated. The principle will be explained based on FIG.

In Fig. 2, TP (where n -1, 2, 3) is n CAD/CA) the hand position commanded on I system 1;
Posture command point, I? P (where n = 1.2.3) indicates the destination point that the hand of the robot 4 reaches by actually operating after being taught by each command point TP on the CAD/CAM I.

Y is the position/orientation vector between the command point TP on the CAD/CA) Il and the command point n1+10 TP, π. , n+1 indicates the reached position/orientation vector between the reached point RP and the reached point RPn+1 of the robot n+1 and the robot 4.

When the robot 4 is actually operated using the command values generated on the CAD/CAM system 1, the hand of the robot 4 will not reach the command point TP in real space due to the origin error Δb.
As shown in Figure 2, the reached position/orientation has a position/orientation error d.
Achievement point 1 shifted by F (however, n = 1, 2.3)? Reach P. At this time, the reached position/orientation vector π in real space is expressed by the following equation: n, n+1.

R' = -d F +F' 10d F'
n+1n, n+1n
n, n+1...(2) However, ``?4m(rxIrylrZ, rφ, rθ1r4)T...
...(3) R'-(R, R, R, Rφ, Ro, R4)Tx
y Z ... (4) Here, (r, r,, r) is a vector indicating the hand position,
(rφ, θ, r4) are vectors indicating the direction of the hand, and Y indicates their composite vector. The same applies to π.

Therefore, π −d F + F + d F 2 ...
...(5) R= dF + F2a+dFa
...=(6) Assuming that the position/orientation error dF' is a minute change due to the origin error Δp, the position/orientation error d
Fn is expressed by the following equation using a Jacobian matrix based on minute rotation.

dr″−J Δb...
...(7) n n n However, Δf=(Δθ, Δθ...Δθ1).

n l 21 is the degree of freedom of the robot 4.

Here, we will take an example of an articulated robot with 6 degrees of freedom (5 degrees of freedom).
), (6), and (7), and the measured position information R, R of the reached position/orientation vector π.

n, n+1
A method of determining the origin error Δb using YR, assuming that the origin error db is equal to Δj, will be explained.

From equations (5), (6), and (7), the above (8)
Expression (9) is expressed as an element as shown in the following expression.

When measuring the hand position of the robot 4 with a three-dimensional measuring device, only the positions of the y-axis, y-axis, and Z-axis are measured, so equation (10) is used.
Since the attitude values shown in the fourth to sixth rows on the right side of equation (11) cannot be obtained, the following equation can be obtained by taking equations (10) and up to the third row of equation (11).

The first to third lines are shown.

Therefore, the equation (13) θ is obtained from the equation (12), and the origin error Δb can be obtained by calculating the Jacobian matrix J.

The Jacobian matrix J0 is the unit rotation axis of the joint based on the robot coordinate system x, y, z in Figure 3, as described in "Robot Control" (published by the Japan Society of Instrumentation and Self-Help Control), pages 22 to 28. Set the vector to F and the position vector from the joint center to the hand of robot 4.

If the minute change of the hand is dF' and the minute rotation angle of the joint is Δb, then the following equation can be obtained from the rotational motion of the rigid body.

However, P Xrf) indicates the cross product, and dφ indicates the minute displacement of the rotation n joint. From these equations (14) and (7), the Jacobian matrix Jθ can be calculated from the rotation axis of the joint of the robot 4, the position vector from the center of the joint to the hand of the robot 4. .

The operation of this embodiment constructed according to the above operating principle will be explained below based on the flowchart shown in FIG.

First, the procedure for determining the origin error Δj will be explained. When the detection start signal for the origin error Δj is input (step 20), the set value '27 of the rotation angle of the joint for each designated point TP commanded on the CAD/CAM system 1 is first set.
(However, -0 for the number of measurement points n-1 to 3 corrected repeated rotations) is obtained (step 21). Origin error Δ at this initial time
Assuming that θ° is zero (Stella ■ 23), the initial setting value for the rotation of the joint. is sent from the boss processor 3 to the robot control means 41 to sequentially move the hand of the robot 4 to each destination aRp corresponding to each instruction point TP on the CAD/CAM system 1 (step 29). As the hand position of the robot 4 moves, the position of each reaching point RP is sequentially measured by the tertiary light measuring device and sent from the input device 5 to the post-processor 3 (step 25).

When measuring the hand position of the robot 4 with this coordinate measuring device, as shown in Fig. 5, the coordinate system of the surface plate 9 on which the coordinate measuring device is installed, that is, the measuring device coordinate system (x, y, z) and The robot base coordinate system (x,
y, ZR) must be parallel.

R Therefore, before actually measuring the hand position of the robot 4 with the three-dimensional jPj measuring device, the installation error of the three-dimensional measuring device is corrected.

This correction is made by first rotating only the first link 11 on the ZR wheel axle of the robot 4, and making P P.

0゛I Measure the coordinate values of three points P.

20'I Identify the plane passing through point P. Next, 3 points P. .

A plane passing through the PP point and the measuring instrument coordinate system (x).

1' 2 Cyo
, z) with the two axes (0, 屹 a), the rotation α around the CC demand.

〒-Trans(0, O, a) Rot+(x, α)
Rot+ (y, β) (15) The inverse matrix 〒-1 of this tilt correction matrix 〒 is expressed in the measuring instrument coordinate system (x
The robot base coordinate system (x, y, z) is multiplied by the measured value of the three-dimensional measuring device
z) to obtain the measured value X based on RI? I can do R.

Each reaching point RP of the hand of robot 4 obtained as above
The position information of each point TP on the CAD/CAM system 1 is sent to the minute relational expression calculation means 32, and
Each designated point TP is stored in the storage means 31 in which position information is stored.
The positional information of is sent to the infinitesimal relational expression calculation means 32, which calculates the infinitesimal relational expression shown in equation (13) using these positional information and the Jacobian matrix θ J , and the arithmetic means 33 calculates the infinitesimal relational expression. The relational expression is solved to find the origin error Δb1 (step 2G).

Since the origin error Δb is assumed to be a minute amount, the above operation may be repeated (step 29) in order to accurately determine the origin error Δb. In other words, the origin error Δb1 calculated as shown in FIG. (step 25),
By repeating the above operation to obtain the origin error Δb2,
A chip 28) in which the determined origin error Δb 5 has reached a convergence boundary value ε, for example, a value at which the hand positioning accuracy of the robot 4 reaches the position reproducibility accuracy of the robot 4.

Next, when actually operating the robot (Step 2)
2) Add the origin error Δp to the command value p.

p, and the robot is operated based on this value (step 30).

Table 1 shows the results of calculating the origin error Δb of each axis by specifically applying this example to a 5-degree-of-freedom articulated robot. In order to calculate the accurate origin error Δb, theoretically, the three points reached by the robot hand, 1? It is sufficient to measure Pn, but Table 1 shows the
The case where calculation is performed based on the data of the destination points RPn (n-1 to 5) is shown.

Table 1 In Table 1, i-1 to i-5 indicate each axis of the 5-degree-of-freedom articulated robot.

Next, the positioning accuracy was corrected using the origin error Δθ shown in Table 1, and the absolute positioning accuracy was measured at arbitrary 18 points within the operating range of the robot 4. The results are shown in FIG. 6th
The figure shows the change in absolute positioning accuracy depending on the number of correction repetitions, with the horizontal axis representing the number of correction repetitions and the vertical axis representing absolute positioning accuracy (m+n). Absolute positioning accuracy with correction is 3.68m+
It was improved from n to 0.51 mm. The positioning repeatability accuracy guaranteed by the robot specifications used for this measurement is ±0.5+nm, so the positioning repeatability accuracy guaranteed by the robot specifications with two correction repetitions is the same as the positioning repeatability guaranteed by the robot specifications. It was possible to obtain absolute positioning accuracy with a degree of accuracy.

Furthermore, in this embodiment, as explained above, when measuring the reach point RPn of the hand of the robot 4, the geometric relationship between the measuring instrument coordinate system and the robot base coordinate system is calculated, and the robot in the robot base coordinate system is Since we are calculating the origin error Δb of 4, after the origin error Δj has been corrected once, there is no need to compensate the robot itself, and when installing the robot, it is only necessary to correct the base error Δπ. ,
Sufficient positioning accuracy can be obtained.

In this embodiment, a case has been described in which the command values for each axis of the robot are given by a CAD/CAM system, but the command value for each axis of the robot may be given by a computer system or the like.

[Effects of the Invention] As explained above, the present invention allows each axis of the robot to be adjusted based on the command values of multiple positions/postures or angles of each axis and the measured values of the relative positions of the robot in the robot coordinate system based on the command values. Since the origin error is calculated and the absolute positioning accuracy of the robot is corrected using this calculated value,
Absolute positioning accuracy can be improved without using detection values from a detector inside the robot.

Furthermore, since absolute positioning errors can be corrected without using a robot internal detector, it is possible to improve the absolute positioning accuracy of any general-purpose robot.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing an embodiment of the present invention, FIG. 2 is an explanatory diagram showing the operating principle of the above embodiment, FIG. 3 is an explanatory diagram of Jacobian matrix calculation in the above embodiment, and FIG. 4 is an explanatory diagram showing the above embodiment. A flowchart showing the operation of the example, FIG. 5 is an explanatory diagram of error correction between the measuring instrument coordinate system and the robot base coordinate system in the above embodiment, FIG. 6 is a change characteristic diagram of absolute positioning accuracy according to the above embodiment, and FIG. 7 FIG. 2 is an explanatory diagram showing error classification of the robot's hand position/posture. 1... CAM/CAM system, 2... Graphic display, 3... Post processor, 4...
Robot, 5... Input device, TPI to TP3...C
Instruction points on the AD/CAM system, RPI to RP3...
・Achievement point of robot minions.

Claims (3)

[Claims] (1) In a robot positioning error correction method in which the position/orientation or the angle of each axis is given as a command value to the robot control device and the robot's position/orientation is controlled, (a) multiple positions/orientations Or, give the command value of the angle of each axis to the robot control device, and (b) actually measure the position of the robot's hand in the robot's base coordinate system corresponding to the command value of each position/posture or the angle of each axis, and ( c). Calculate the origin error of each axis of the robot from the relative values of the command values for each position/posture or angle of each axis and the actual measured values of the hand position using a micro-relationship equation. (d). A method for correcting positioning errors in a robot, characterized in that the command value of the posture or the angle of each axis is corrected by the origin error of each axis to control the position and posture of the robot. (2) The relative position measurement of the robot in real space taught by command values is performed by correcting the robot base coordinate system and the coordinate system of the position measuring device so that they are parallel to each other. robot positioning error correction method. (3) A robot positioning error correction method according to claim 1 or 2, wherein the positioning error in real space is repeatedly corrected based on the origin error of each axis calculated from the minute relational expression.