JPH01121904A - Correcting method for positioning data - Google Patents

Abstract

PURPOSE:To shorten the processing time and to accurately position a robot, etc., by correcting and converting arbitrary data on a coordinate system on a design to position data of an actual coordinate system based on relations between the coordinate system on the design of a positioning device of the robot, etc., and the actual coordinate system of the positioning device. CONSTITUTION:An object to be moved is moved to three point to measure position data expressed with the internal coordinate system of these three points which determines a plane with respect to a reference material having the plane as the reference. Based on coordinate data expressed with an external relative coordinate system of three points and position data of three points on the internal reference coordinate system obtained by measurement, a conversion matrix from the external relative coordinate system to the internal reference coordinate system is obtained by an arithmetic process. Arbitrary position data on the external relative coordinate system is converted to position data on the internal reference coordinate system by the conversion matrix, and position data expressed with the external relative position coordinate system is corrected to that on the internal reference coordinate system.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention is aimed at reducing coordinate errors between a designed positioning position and an actual positioning position in a device such as a robot that requires a positioning operation. The present invention relates to a method for correcting positioning data for correcting.

[Prior Art] For example, in devices such as robots that require positioning operations, teaching is generally performed to teach position coordinates. This teaching includes so-called online teaching, in which teaching points are directly taught one by one using a device such as a robot, and offline teaching, in which teaching points are taught indirectly from design data. Although the former online teaching has the advantage of high teaching accuracy, it takes a huge amount of time to teach all the teaching points and is not efficient. Additionally, there is a safety issue because the operator enters the operating area of the robot. For this reason, offline teaching has recently been attracting attention.

For offline teaching, for example, Japanese Patent Publication No. 62
-108314, JP-A No. 62-139003, etc.

[Problems to be Solved by the Invention] However, as shown in FIG. 18, in the design of positioning devices such as robots, the coordinate axes of the positioning system are, for example, orthogonal three-dimensional coordinate axes (x, y, z, etc.).

Although it can be designed to exactly match the coordinate system (X) of the device when it is actually manufactured and assembled.

y, z, e), there are always errors.

Therefore, when off-line teaching of teaching points is performed to a robot device based on design data, a situation may occur in which the finger does not move to the expected position even when the robot device is operated.

Therefore, the present invention was proposed in order to eliminate the drawbacks of the above-mentioned conventional examples, and its purpose is to improve the design of a positioning device that moves a moved part eight times at a predetermined position taught in advance. Based on the relationship between the above coordinate system and the coordinates of the actual positioning device, any position data based on the designed coordinate system is corrected and converted into position data based on the actual coordinate system. The aim is to propose a method for correcting positioning data.

[Means and effects for solving the problem] One configuration of the present invention for achieving the above-mentioned problem is to have an internal reference coordinate system, and to move the object to be moved to the position represented by the internal reference coordinate system. In a device having a positioning part for moving a part, regarding a reference object having a plane serving as a reference,
a measuring step of measuring position data represented by the internal reference coordinate system at three points that determine the plane by moving the moved part to these three points; and a calculation step of calculating a transformation matrix from the external relative coordinate system to the internal reference coordinate system based on the coordinate data represented by the coordinate system and the position data of the three points according to the internal reference coordinate system determined by the measurement; , a conversion step of converting arbitrary position data based on the external relative coordinate system into position data based on the internal reference coordinate system using the conversion matrix, thereby converting the position data represented by the external relative coordinate system into the position data expressed in the internal reference coordinate system. It is characterized by correcting the position data by the system.

Another configuration according to the present invention for achieving the same purpose also includes a three-dimensional internal reference coordinate system containing rotational angular coordinates around a rotation center, and a position represented by the internal reference coordinate system. In a device having a positioning unit for moving a moved part, position data expressed by the internal reference coordinate system of three points that determine the plane with respect to a reference object having a plane as a reference, A first measurement step of measuring by moving the moved part to these three points while maintaining a constant rotation angle; a second measuring step of measuring position data represented by the internal reference coordinate system of at least one of the three points by moving the point while maintaining a rotation angle with a predetermined phase deviation; , the coordinate data of the three points expressed by an external relative coordinate system, which is a three-dimensional relative coordinate system including rotation angle coordinates around the rotation center, and the coordinate data of the three points determined by the first and second measurements. a calculation step of calculating a transformation matrix from the external relative coordinate system to the internal reference coordinate system based on the position data of the three points according to the internal reference coordinate system, and converting arbitrary position data according to the external relative coordinate system into the transformation matrix; and a conversion step of converting the position data into the position data according to the internal reference coordinate system, thereby correcting the position data represented by the external relative coordinate system to the position data according to the internal reference coordinate system.

[Examples] Examples according to the present invention will be described below with reference to the accompanying drawings. An embodiment to which the correction method of the present invention described below is applied is a system for off-line teaching of teaching points for a robot assembly device.

<Robot Assembly Device> First, the appearance of the robot to which this offline teaching system is applied will be explained with reference to FIG. The robot assembly device 1000 shown in the figure is a well-known so-called XY type robot, in which a servo motor M□ moves an arm 502 in the Y direction.
MB2 moves arm 501 in the X-axis direction, MB2 moves arm 1
in the θ direction (rotation), and MB4 moves the arm 1 in the Z direction. A force sensor 2 and a finger 3 are attached to the tip of the arm 1 via a holder 4. The fingers 3 grasp the parts in the pallet 100 placed on the table 500 one by one and finish them into one product on the assembly stage 504. Incidentally, 503 is for hanging a spare finger when a plurality of fingers are required to assemble one product. The controller of this robot is stored in a location not shown in FIG. 2, and is basically equivalent to a well-known controller.

600 is an offline teaching controller for performing offline teaching to the robot 1000, and in reality, a so-called personal computer is used.

Figure 3 shows pallet 1 used in this robot 1000.
00 appearance is shown. This pallet 100 has predetermined external dimensions, and a plurality of necessary parts are stored therein. FIG. 4 shows an example of the storage state of parts. In the example of FIG. 4, the types of parts in one pallet are the same, so the fingers 3 pick up the parts one by one from the pallet 100 on the stand 500. When taking out parts in order, it is possible to point to the part storage positions on the pallet that are partitioned and numbered in the shape of grids on the base. That is, the finger 3 grips a component, which means that the gripping portion of the finger 3 matches the gripping position of each component (X).

y, z, e) from the robot controller to the robot.

In the example shown in Fig. 4, it was assumed that all parts in one pallet were the same, but this does not apply to the present invention, which is based on the fact that different parts are housed in a pallet rather than wood. You can. The point is that the gripping position at the storage position of the component within the pallet can be used as a teaching point for teaching. Although it is possible to give the coordinate values of the part gripping position within the pallet as a teaching point when placed on the stand 500 on the design drawing, when it is operated with the actual robot 1000, it may be misplaced. The movement of the finger 3 in the direction has been explained in connection with FIG. 18.

<Principle of teaching point correction> Therefore, the design values (ideal coordinate system (x, y).

2, θ)), the actual coordinate system of the robot (X, Y
, Z, and l need to be known to what extent they are deformed. That is, how and to what extent should the teaching point based on (x+y, z*θ) be corrected to obtain an accurate teaching point. According to the offline teaching of this embodiment, the component gripping position (Q+,
Q! ...) design teaching points (x, y
,z.

θ) is given as a corrected teaching point (x, y, z, e) as shown in FIG. It moves according to the robot coordinate system as if it were a robot.

Furthermore, according to this embodiment, once the teaching point correction method becomes clear, it will also become clear that this correction method provides a fast and efficient offline teaching method.

Now, FIG. 5 shows a member 20 that serves as a reference for this correction. This member is shown in Fig. 5 as having the same external dimensions as the pallet in Fig. 4, but if the external dimensions and the dimensions of the three points R1-R3 are known, it is a special member. It doesn't matter if it isn't. This member 20 will be referred to as r reference pallet J hereinafter. This reference pallet 20 is placed on a stand 500 in FIG.
is placed in the same way.

If the internal coordinate system of the robot assembly device 1000 (hereinafter referred to as r-robot coordinate system 1) is translated or sheared one rotation with respect to the designed coordinate system (hereinafter referred to as r-relative coordinate system 1), If the relationship is an affine transformation such as expansion (such an assumption is a reasonable assumption if the robot arm does not bend, etc.), then this transformation matrix (hereinafter, this matrix will be specifically If possible, for the design teaching points of the pallet 100 used in the actual assembly shown in FIG. 4 (all of which can be accurately known in advance),
By applying the transformation matrix of the affine transformation, teaching points based on the robot coordinate system can be obtained.

Keeping in mind that the plane remains a plane even after affine transformation, three points on the plane of the reference palette 20 (points R1+R2+R3) shown in FIG.
) with the finger 3 of the robot device 1000 itself, and position data of those points according to the robot coordinate system, (XR1, YRl, ZRI) ~ (XR31YRl,
XR3), position data based on the relative coordinate system of those three points (XRII V*rr ZRI) A-(XR31yn
Since s+XR3) is known, it is possible to obtain the transformation matrix of the affine transformation.

It is important to detect whether the finger 3 has accurately pointed at the three points, and this is achieved by using the force sensor 2 in this embodiment.

The above discussion is based on the design ideal coordinate system (x.

y, z, θ) versus the actual coordinate system (X, Y.

The extent to which z and e) are deformed is determined by coordinate transformation as shown in FIG. If you do not rotate the finger you are using and keep θ constant, you do not need the rotational coordinate value e to determine the teaching point, so the position according to the robot coordinate system obtained by the three-point measurement Information (XR1, YRl, ZRI) ~ (XR3,
YRl, XR3) is sufficient information to find the teaching point in the robot coordinate system. However, in a robot assembly device or the like, in addition to simply the xYz coordinate data, the amount of rotation e around the Z axis is important as a teaching point. Generally, the fingers attached to the robot are offset with respect to the center of rotation around the Z-axis, so unless e is determined, it is not possible to point the finger 3 to the teaching point. By the way, when the finger is attached to the arm, just like in the case of a robot's arm drive system,
There will be an error in the angle when the finger is attached in the design, but there will be an error in the angle when the finger is actually attached. Therefore, finding this error is essential for off-line teaching. As will be described in detail later, the installation error is determined by finding a vector connecting at least two or more finger bins (A and B in Figure 1) in the relative coordinate system and the robot coordinate system, and then using these two vectors (P The intersection angle between the vectors is determined from the inner product of P' and P', and the position is determined from the position vector.

To summarize the above discussion, the off-teaching method of this embodiment is shown in FIG. In Figure 8, the horizontal direction shows the flow of a sequence along the flow of time, and the vertical direction shows the flow of data. 0, the coordinate values of the three reference points R1 to R8 in the robot coordinate system are measured by a force sensor, and based on these coordinate data and the coordinate data of the above three points in the relative coordinate system, the robot is adjusted from the relative coordinate system. Find the transformation matrix H to the coordinate system.

Next, when attaching the fingers, the error angle αO is determined from the inner product of vectors obtained from the coordinate values of the finger bins A and B based on the robot coordinate system and the relative coordinate system.

Next, the data (x, y, z, θ) in the relative coordinate system of each teaching point of the palette actually used (for example, the one shown in Figure 4) is multiplied by the affine transformation matrix H, and ( X, Y, Z, e')-H(x,
y, z, θ)...(1), and then θ=e' + α0...(2), and finally (x,
y, z, e).

Offline teaching will be explained in more detail below.

<Offline Teaching System> FIG. 1 shows the configuration of an offline teaching system according to this embodiment. Arm 1 etc. are the same as shown in FIG. 4 is a holder for fixing the force sensor 2 to the finger 3. A pair of finger pins A and B are attached to the finger 3. The force sensor 2 is
For example, it is a 6-axis force sensor using a strain gauge.

The offline teaching controller 600 is connected to the force sensor 2 and the robot controller 10.

That is, the output from the strain gauge of the force sensor 2 passes through the amplifier 7, is converted into a digital value by the A/D converter 8, and is converted into a digital value by the A/D converter 8.
The data is stored in the AMII and processed by the CPU 9. The CPU 9 is controlled according to an offline teaching program stored in the ROM 12 or the like. As will be described later, this program controls the arms 501, 502 .
Inching the 1st class. This inching operation moves the finger pins to the reference hole positions of the reference pallet 20.

Further, the program monitors the output of the force sensor 2, which changes according to the inching operation, and detects that the finger bin has reached the center of the reference hole.

<Detection of the center of the reference hole> In principle, it is not difficult to accurately measure the coordinate values of the reference point shown in FIG. 5 according to the robot coordinate system. For example, a fine-tipped probe (not shown) may be provided at the tip of the finger, and changes in the output of the force sensor 2 may be monitored when the probe is positioned at the reference point. However, although these special probes can accurately detect the position, they have a different shape from the fingers that are actually used, and therefore offline teaching cannot be performed unless the difference in shape between the probe and the actual finger is corrected. It is impossible to do so, and such a troublesome operation is inefficient. Therefore, in this embodiment, the fingers used in actual assembly are used for detecting the reference point position, and for this purpose, the reference points R, -R, in FIG. The hole is cylindrical and is large enough to fit the tip of the hole inside. The advantage of using offset finger pins as shown in FIG.
It is possible to obtain the coordinate values of the center of the reference hole.

So, how can we use such pins and reference holes?
Explain how to find the reference point (which can be defined as the point located at the center and bottom of the reference hole).

As mentioned above, the finger pin is inserted into the reference hole by inching. Then, as shown in FIG. 9A, the finger pin A (B) contacts the wall of the hole by an inching operation while keeping the angle in the e direction constant and monitoring the output of the force sensor 2. Check location. Similarly, in the Y direction, contact with both walls is detected while moving the finger pin as shown in FIG. 9B. For this detection, the output of the force sensor is shown in Figure 10A~
As shown in FIG. 10B, this can be seen from the fact that it changes significantly upon contact with the wall. Therefore, the average value of the two coordinate values in the robot coordinate system at the time of contact with the hole wall gives the coordinate value of the reference point at the center of the hole.

Regarding the Z direction, when the finger pin is moved in the X direction and comes into contact with the wall, the finger pin is lowered downward. Then, collect the Z coordinate value at the time of contact with the bottom. Similarly for the Y direction, the Z coordinate value of the reference point is determined from the average value of the total four Z coordinate values.

In this way, when the tip of the finger pin is positioned at the reference point R with a constant rotation angle e, the coordinate value in the robot coordinate system of the rotation center on the Z axis (see FIG. 1) is given. Figure 9 like this.

The method for measuring coordinate values shown in FIG. 10 is not only used for offline teaching in this embodiment, but also for general
This method is suitable for determining coordinate values when the object to be measured is tilted.

<Determining the transformation matrix> As mentioned above, the plane is determined by determining the three points, so from the position coordinates of the three points in the relative coordinate system and the robot coordinate system, the transformation matrix H between these two coordinate systems can be determined. Desired. FIGS. 11A and 11B specifically represent the discrepancy between the relative coordinate system and the robot coordinate system, with the former representing a relationship of parallel movement, and the latter representing a relationship in which shear is added to the parallel movement. For parallel movement as shown in Figure 11A, ΔX and ΔY can be easily determined from the difference in the coordinate values of the three points, and it is easy to obtain the correct teaching point from this correction value by off-line teaching. .

In the case as shown in Fig. 11B, a combination of parallel movement processing as shown in Fig. 19A, shearing processing as shown in Fig. 19B, and rotation processing around the Z axis as shown in Fig. 19F is used. If we know the matrix elements shown in these figures,
First, if a parallel movement is performed as shown in FIG. ttC, a rotation process is performed as shown in FIG. 11D, and then a shear process is performed, the two will overlap as shown in FIG. 11E. Therefore, if such a transformation matrix is found, the teaching point in the robot coordinate system can be easily found.

FIG. 12A is a flowchart of a control program for calculating such a conversion matrix. In step S2, the rotation angle e of the arm 1 is set to an arbitrary offset angle θ0. In step S4, the finger pin A is moved to one reference hole position by inching,
In step S6, the center coordinates of the reference hole are determined by a method such as that shown in FIG. 9 described above. 13A and 14A show the state when the finger pin A points to the center of the reference hole R1. The coordinate values in the robot coordinate system of the rotation center at this time are (XRIA, YR
IA, ZRIA, eo).

Next, in step S8, the rotation angle θ is set to (eo+180) degrees. That is, the arm 1 is further rotated by a half turn from the position eo. Next, in step S10, the coordinate values of the center of rotation when the finger bin A is placed at the center of the reference hole R are determined. The coordinate values obtained at this time are (X'RIA, Y"RIAI Z'RIAI eo"
180). This state is shown in FIGS. 13B and 14B. Thus, even when the finger A is offset from the rotation center, the coordinate values in the robot coordinate system when the rotation center points to the reference hole R1 in step S12 are obtained as follows (3) . That is, as is clear from FIG. 14C, which corresponds to FIG. 13C, when the above coordinate values are given as teaching points, the robot controller 10 controls the movement of the robot arm so that the center of rotation is on the reference hole R1.
is controlled by The operations from step S2 to step S12 are performed for all three points (R1 to Rs) on the reference pallet 20. The coordinate values obtained by searching the reference points R2 and R3 with finger pin A at angle e0 are (XR2A * Y R2A + Z R2A + e
O) (XR3A + YIIIA + ZR3A r
ea) and (X'112A, Y"R2A, Z'112A, e
at 80 ) (X'R3A+ Y'R3A+ Z
'R3A* θo+ 1 ao). Furthermore, when the center of rotation is above the point R,, R, the following robot coordinate values are obtained: (4) (5). Regarding R2 and R3, the coordinate values are calculated in the state of eo, and the eo of finger pin A is
From the relationship between the coordinate value of R+ and equation (3), the same result can be obtained even if it is calculated by parallel movement.

Next, in step S16, the coordinate values on the relative coordinate system when the center of rotation is located at the designed points R8 to R3 are manually entered into the RAM from the outside. Such a value is on the order of 500 (
These are the design coordinate values when the reference pallet 20 is placed on FIG. 2). This is because the designed position data of the stand 500 is known.

Therefore, the coordinate values of these reference points R1-R5 on the relative coordinate system are (XRII yRll Z*1+eo)...
...(6)(X R2+ yR2, Z R2+ eo
) ”” (7) (X R3+ '/ R3+ Z
R3 + θ0)... (8) If these are the vectors from the origin of the coordinate axes to R1 to R3, then, for example, the vector in equation (6) becomes the vector in equation (3). Since this is nothing but an affine transformation as shown in FIG. 19, the matrix elements of the above transformation can be calculated inversely. That is, the element calculation of the conversion matrix in step S18 is to solve the equation from (X, Y, Z, θ) "H" (x, y, Z, θ) to obtain the coefficients of the matrix.

Finger attachment is an angle measurement> If the robot arm 1 does not rotate at all around its axis, the data necessary for off-line teaching will be available by obtaining the above conversion matrix.

This is because if the arm 1 does not rotate, the angle of attachment of the fingers does not matter. However, in reality, in many robot devices, the arm rotates in the θ direction, so
Teaching the rotation angle is also important.

Therefore, according to FIG. 12B, the control program for this installation angle calculation will be explained. First, in step S20, the rotation angle e is set to eo, and the finger pin B is inserted into the reference hole R1.
I'll take it to In step S24, coordinate values of the rotation center when the finger bin is located at the center of the reference hole are determined.

This state is shown in FIGS. 13D and 14D. The coordinates of the rotation center at this time are (XRIB, YRI!l, ZRIIS, e
o ), vectors P and P in FIG. 15, which shows the states of FIGS. 14A and 14D superimposed.

The intersection angle between (=P) and the coordinate axis represents the attachment angle of the finger 3 in the robot coordinate system. The vectors P and P have the following values in the robot coordinate system. Based on the same idea, two points (X RIA + '/RIA + zRIA +
θ0) (X RIB + '/ RIB + Z R1
[1r 00) is connected to the vector P'IP'2(=P'
) is expressed as, and HP', which is obtained by multiplying this P' by the transformation matrix H obtained above, is the vector P in the robot coordinate system.
However, due to manufacturing errors and mounting errors of the fingers 3, P ≠ HP'. By the way, the above-mentioned errors appear only as errors around the Z-axis of the arm 1. Therefore, as shown in FIGS. 16A and 16B, if the error angle is α0, (P, HP')=1PIIHP'1cosa.

It is.

By the way, (P, HP') is expressed by the inner product of the coordinate values and can be calculated. Also, IP+ and IHP
'1 is known, so from the above equation, the error angle α for the installation of finger 3 is given. is determined in step S26 as .

Note that the program shown in FIG. 12B is applied to all fingers actually used in the robot device 1000.

<Teaching point data generation> This data creation procedure is shown in FIG. 12C.

Since the relative positional relationship between the reference pallet and the pallet actually used is given in advance as the distance in the x, y, and z directions between the reference point R and the actual part gripping position as shown in FIG. Such position data for the state in which the camera is placed is inputted in step S30. In step 332, (
1) Multiply the conversion matrix obtained in FIG. 12A as shown in the equation. In step S34, the attachment of the finger determined by the program shown in FIG. 12B has an angular error α. Therefore, e is corrected according to equation (2).

In step 336, the teaching point (x, y, z, e) corrected in this way is transmitted to the robot controller 1.
Output to 0.

In a loop from step 332 to step S38, all teaching points of one pallet are found and corrected. When finished for one pallet, step S30
~ In the loop of step S40, teaching points for all fingers used in this robot device are calculated and sent to the robot controller 10. The transformation matrix obtained based on one reference pallet can be applied to other fingers and other pallets, as long as the same robot device is used, the correspondence between the robot coordinate system and the relative coordinate system is unique. This is because it is relevant.

<Effects of the Example> In this way, the robot controller has been taught the teaching points regarding all the parts and fingers necessary for assembling one product. The teaching points are also corrected for intersections specific to the robot assembly device, and are accurate. Further, by using the conversion matrix obtained from the measurement of one reference pallet, all the teaching points of the finger gripping positions of the pallet to be actually used can be calculated at high speed.

[Effects of the Invention] As explained above, according to one configuration of the positioning data correction method of the present invention, designed position data is corrected to positioning data that accurately corresponds to an actual positioning device.

According to another configuration, in addition to the effects of the invention described above, it is also possible to correct the rotational coordinate.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of an offline teaching system when the present invention is applied to the system, FIG. 2 is a perspective view of a robot assembly device in which the offline teaching system of the above embodiment is used, and FIG. 3 is a diagram of the system shown in FIG. A perspective view of a parts storage pallet used in a robot assembly device, Fig. 4 is a plan view showing one aspect of the pallet, Fig. 5 is a plan view of a reference pallet, and Fig. 6 is a design teaching data and after correction. FIG. 7 is a diagram showing the relationship between the robot coordinate system and the relative coordinate system of the reference pallet. FIG. 8 is a diagram explaining the schematic sequence of the offline teaching method of the embodiment. FIG. 9A , FIG. 9B, FIG. 10A, and FIG. 10B are diagrams explaining how to find the center position of the reference hole, and FIGS. 11A to 11.
Figure E is a diagram explaining the concept of matching coordinate systems, Figures 12A to 12C are flowcharts of programs related to offline teaching control, Figures 13A to 13C, and Figures 14A to 14C. 13D, 14D, N15, 16A, 1
Figure 6B is a diagram explaining the method for determining the angular error when installing fingers, Figure 17 is a diagram explaining the relative positional relationship between the reference pallet and the pallet used, and Figure 18 is the design position and actual position. Figures 19A to 19G are diagrams explaining the conversion matrix. In the figure, 1...Z-axis arm, 2...force sensor, 3...finger, 4...holder, 7...amplifier, 8...
・A/D converter, 9...CPU, 10...Robot controller, 11 ・RAM, 12・ROM, 1
3 - Keyboard, 20... Reference palette, 100
...Pallet used, 500...Pallet stand, 501
-...X glaze arm, 502...Y axis arm, 503...
... Spare finger suspension stand, 504... Assembly stand, 60
0...Offline teaching controller, 100
0...Robot assembly device, A, B...Finger pin, MRI~MR3 river servo motor, R1-R3...
It is a reference point (hole). Figure 1 Figure 2 Figure 4 Figure 7 Figure 9A Figure 98 Y Figure 108 Figure 11A Figure 11D Figure 11E Figure 512A θ=00 Pe1 (XRIA, YRIA, ZRIA, eo) No. 134
Figure θ=eo+180' (n(p+a, Yiq+a, Zi+a, θo+l80
) Figure 13B θ=θ. Figure 13D -α Fig. 15 Fig. 17 Fig. 16A S Fig. 168 Fig. 18 Rotation around b Rotation around the y-axis Rotation around the Z-axis Figure 19F Figure 19G

Claims (6)

[Claims] (1) In an apparatus having an internal reference coordinate system and a positioning section for moving a moved part to a position represented by the internal reference coordinate system, a reference having a plane serving as a reference. a measuring step of measuring position data of an object expressed by the internal reference coordinate system at three points that determine the plane by moving the moved part to the three points; and the three points. Based on the coordinate data expressed by the external relative coordinate system of the position and the position data of the three points according to the internal reference coordinate system determined by the measurement, convert from the external relative coordinate system to the internal reference coordinate system. By comprising a calculation step for obtaining a conversion matrix, and a conversion step for converting arbitrary position data based on the external relative coordinate system into position data based on the internal reference coordinate system using the conversion matrix, A method for correcting positioning data, comprising correcting expressed position data to position data based on the internal reference coordinate system. (2) In the measuring step, the position is measured using a probe provided at the tip of the moved part and connected to a force sensor. The method for correcting positioning data described in section. (3) The three points on the reference object are cylindrical holes having a predetermined depth and radius, and the measurement step includes a step of moving the probe into the hole, and a step of moving the probe into the hole. a step of monitoring the output of the force sensor and collecting a plurality of coordinate data of the reference coordinate system at a position of change in the output of the force sensor while moving so as to make contact with a wall surface of the force sensor; and calculating coordinate data of the center of movement of the moved part based on the data, and the coordinate data of the center of movement is used as position data based on the reference coordinate system of the three points. The positioning data correction method described in Section 2. (4) Contains rotation angle coordinates around the rotation center inside 3
In a device having a dimensional internal reference coordinate system and a positioning unit that moves a moved part to a position represented by the internal reference coordinate system, regarding a reference object having a plane serving as a reference. , the positional data represented by the internal reference coordinate system at three points that determine the plane is measured by moving the moved part to these three point positions while maintaining a constant rotation angle; 1 measurement process and moving the moved part to at least one of these three positions while maintaining a rotation angle that is the constant rotation angle and a predetermined phase deviation. a second measuring step of measuring position data represented by said internal reference coordinate system of at least one of the points; and an external relative coordinate system including rotational angular coordinates about said center of rotation. The external relative By comprising a calculation step for obtaining a transformation matrix from a coordinate system to an internal reference coordinate system, and a conversion step for converting arbitrary position data based on the external relative coordinate system into position data based on the internal reference coordinate system using the conversion matrix. . A method for correcting positioning data, comprising correcting position data expressed by an external relative coordinate system to position data based on the internal reference coordinate system. (5) The positioning data correction method according to claim 4, wherein the predetermined phase deviation is 180 degrees. (6) First and second probes offset from the rotational axis are provided at the tip of the moved part, and these probes are connected to a force sensor. are opposed to each other around the rotation axis at a predetermined phase angle, the three points on the reference object are cylindrical holes having a predetermined depth and radius, and the first and second measurement steps is a step of moving the first probe into the hole, and monitoring the output of the force sensor while moving the first probe in the hole so as to come into contact with the wall surface of the hole; The method includes the steps of collecting a plurality of coordinate data of the reference coordinate system at the sensor output change position, and calculating coordinate data of the movement center of the moved part based on the plurality of coordinate data, and further, Claim 2, further comprising a third measuring step of measuring position data of any one of the three holes while the second probe is rotated by the predetermined phase angle. The positioning data correction method described in .