JP2787891B2 - Automatic teaching device for laser robot - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an automatic teaching device for a laser robot for automatically determining and registering the position of a processing point and the attitude of a processing nozzle when teaching a work having an arbitrary shape with a laser robot. It is about.

[0002]

2. Description of the Related Art In a conventional laser robot, a worker finally irradiates a laser beam to a point to be processed on a workpiece having an arbitrary shape by using an operation button for each axis of the laser robot. The position and orientation of the processing nozzle are determined and registered. Next, by performing playback based on the registered position / posture and processing conditions, laser processing such as laser welding or laser cutting is performed. At that time, in order to obtain a good processing result, it is necessary that the laser beam emitted from the processing nozzle be incident along the normal vector of the processed workpiece. For this reason, the teaching operation of operating the operation buttons of each axis of the laser robot has been difficult and requires the skill and labor of the operator.

[0003] Conventionally, examples of automated teaching work performed by a worker using a three-dimensional laser beam machine or an industrial robot include:
JP-A-62-15063, JP-A-62-4940
No. 5, JP-A-57-178683 and JP-A-3-55184 are known. In these devices, at least three points on a processing work are measured for distance, the position and normal direction of the processing point are determined based on the distance data, and the position / posture of the processing head or tool tip is registered as teaching data. I was

FIG. 8 is a conceptual diagram when teaching a laser robot in the above-mentioned publications and the like. In FIG. 8, 1 is a work to be machined, P ₀ is a point to be machined, P ₁ , P ₂ , P ₃ are points on the work 1 calculated based on information from the distance sensor, and a vector V ₁ is 3 above
This is a normal vector calculated from the points (P _{1 to} P ₃ ).

Next, the operation will be described. The orientation of the teaching data, determined by from points P _1, P _2, P ₃ and normal vector V ₁ calculated by the following equation, superimposes the normal vector V ₁ in the posture of the processing nozzle or tool tip Was.

[0006]

(Equation 1)

[0007] Regarding the position of the teaching data, Japanese Unexamined Patent Publication No. Sho 62-15063 discloses a point P ₁ ,
The average of the respective distances of P ₂ and P _{3 from} the processing head is obtained, and the distances are corrected and determined so as to be a preset distance. Also, determine the distance between the plane and the tool tip is determined from the points P _1, P _2, P ₃ in JP 62-49405 Patent Publication, corrected such that their distance is preset distance It is determined. With these operations, the position and orientation are registered as teaching data.

[0008] In operation method of the conventional teaching data, the position and orientation with respect to the processing point P ₄ having a predetermined curvature on the workpiece 1 shown in FIG. 8, similarly to when workpiece 1 is flat, the distance sensor Are set by three peripheral points (P ₅ , P ₆ , P ₇ ) calculated from. That is, the normal vector of the machining point P ₄ is the point P _5, P _6, P
Since the approximate plane H ₁ as determined from ₇ a normal vector V ₂ which is calculated on the basis, rather than the normal vector V ₃ to the actual curved, As for the position of the machining point P _4, the approximate plane H It was determined by correcting the position above ₁ . Therefore, with respect to the machining point P ₄ on the curved surface, it was not possible to teach the precise position and orientation.
Further, as described above, since it is necessary to arrange at least three distance sensors, their wiring processing is extremely complicated.

Therefore, the present invention has been made to solve the problem of the teaching, and a single sensor disposed at a processing nozzle of a laser robot makes it possible to distinguish a processed workpiece even if it is a flat or curved surface. It is an object of the present invention to provide an automatic teaching device for a laser robot that automatically determines the position and orientation of a processing nozzle automatically and registers the information as teaching data.

[0010]

According to a first aspect of the present invention, there is provided an automatic teaching apparatus for a laser robot, wherein a processing head for irradiating a laser beam guided from a laser oscillator to a processing work having an arbitrary plane or curved surface in an arbitrary direction. One non-contact type distance sensor attached to the processing head, which is movable to a processing point on the processing work,
Detection data conversion means for converting analog data from the distance sensor into digital data, distance calculation means for calculating the distance between the tip of the processing head and the work based on the digital data, and the distance set in advance. Remaining distance calculation means for calculating a difference from the distance required for proper processing, processing nozzle direction calculation means for calculating a movement command for bringing the processing nozzle closer to the processing work by the difference, and Driving means for driving a servo motor of a laser robot, and a shape for calculating a shape data group by rotating the processing head at a preset control point after the processing head reaches a distance based on the movement command. Sampling means; planar curved surface determining means for determining whether the shape data group is a plane or a curved surface; A normal vector group calculating means for selecting five representative points from the shape data group and calculating a normal vector group for the processed workpiece, and calculating an average vector of the normal vector group, Normal vector averaging calculating means for setting a normal vector of a point; tangent calculating means for calculating a tangent from the shape data group and the processing point when it is determined that the surface is a curved surface; A normal vector calculating means for calculating the normal vector of the laser robot, and a position corresponding to the distance set by the driving means as a position of the laser robot at the processing point, and the shape data group is determined based on whether the shape data group is a plane or a curved surface. The position where the calculated normal vector is registered as the posture of the laser robot at the processing point.
Posture registration means.

According to a second aspect of the present invention, in the automatic teaching apparatus for a laser robot, the shape sampling means of the first aspect is configured such that the shape of the processing head is set to a predetermined width along a designated coordinate axis of the laser robot coordinate system. The object is moved in parallel to calculate a group of shape data.

According to a third aspect of the present invention, there is provided an automatic teaching apparatus for a laser robot, wherein the distance sensor according to the first aspect is a contact type, and the shape sampling means is configured to transfer the processing head having the distance sensor to a laser robot coordinate system. Is moved along a combination of parallel movement of a preset width along the designated coordinate axis and movement of the preset distance in the direction of the processing nozzle to calculate a group of shape data.

[0013]

According to the first aspect, it is possible to determine whether the surface is a curved surface or a flat surface based on a sampled data group. Can be. Also, the wiring processing can be greatly simplified because only one non-contact type distance sensor is arranged on the processing nozzle.

According to a second aspect of the present invention, in addition to the function of the first aspect, the shape / sampling means for accurately registering the position and orientation as the teaching data even for a work having a small curvature that cannot be sampled. Can be.

According to a third aspect of the present invention, in addition to the function of the first aspect, since only one contact-type distance sensor is arranged on the processing nozzle, the material cannot be sampled by a non-contact distance sensor. For example, for a work piece such as wood,
Posture can be registered accurately.

[0016]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described below with reference to specific embodiments. Embodiment 1 FIG. 1 is a block diagram showing the entire configuration of an automatic teaching device for a laser robot according to a first embodiment of the present invention, and FIG. 2 is a block diagram of the automatic teaching device for a laser robot according to the first embodiment of the present invention. FIG. 3 is an explanatory diagram showing a relationship between a laser robot coordinate system and a processing nozzle coordinate system. This embodiment corresponds to the first embodiment. In FIG. 1, 1 is a processed work, 2 is detection data conversion means, 3 is distance calculation means, 4
Is a remaining distance calculating means, 5 is a processing nozzle direction calculating means, 6 is a driving means, 7 is a shape sampling means, 8 is a plane curved surface determining means, 9 is a normal vector group calculating means, 10 is a normal vector averaging calculating means. , 11 are tangent calculation means, 12 is normal vector calculation means, 13 is position / posture registration means, 14 is a laser robot, 15 is a processing nozzle attached to the tip of the laser robot 14, and 16 is superimposed on the laser beam . The visible light 17 is a non-contact type distance sensor such as a capacitance type sensor attached to the processing nozzle 15.

Next, the operation will be described. First,
By operating each joint axis of the laser robot 14 with a teaching pendant (controller) (not shown), the position and orientation of the processing nozzle 15 are roughly set with respect to the processing point on the processing work 1. At this time, the visible light 16 projected from the tip of the processing nozzle 15 is adjusted to the processing point. Next, press the automatic teaching start button provided on the teaching pendant,
Start teaching operation.

The detection data conversion means 2 converts the analog data output detected by the distance sensor 17 attached to the tip of the processing nozzle 15 of the laser robot 14 into digital data. In the following distance calculating means 3, based on the digital data converted from the detection data converting means 2, the distance between the tip of the processing nozzle 15 shown in FIG. 1 is calculated.

Next, the remaining distance calculating means 4 calculates a difference between the calculated distance 1 and a predetermined distance L for realizing proper processing, and outputs the difference as a remaining distance Δd of the following equation. . Δd = 1−L (2)

The next processing nozzle direction calculation means 5 calculates a command to move the processing nozzle 15 in the axial direction based on the remaining distance Δd calculated by the equation (2). In other words, for moving the machining nozzle 15 to the Y _N axis direction of the processing nozzle coordinate system 18 shown in FIG. 2 (the axial direction of the processing nozzle 15) by the remaining distance [Delta] d, is decomposed for each moving amount of each joint axis, movement command Is output as This conversion is usually performed using the following homogeneous conversion matrix. The homogeneous transformation matrix Π ₀ located at a distance 1 from the work 1 is represented by the following equation.

[0021]

(Equation 2)

Vector R ₀ = (vector X _N vector Y _N vector Z _N ) where vector X _N is the direction cosine of the X _N axis of the processing nozzle coordinate system 18 with respect to the laser robot coordinate system 19, and vector Y _N is the processing nozzle coordinate system 18 of the Y _N axis laser robot coordinate system 19 the direction cosines for the (indicating the posture of the processing nozzle 15), the vector Z _N is the processing nozzle coordinate system 18 Z _N
The direction cosine of the axis with respect to the laser robot coordinate system 19. Vector P ₀ = (X ₀ , Y ₀ , Z ₀ ) ^t This vector P ₀ is the origin O _N of the machining nozzle coordinate system 18.
Indicates the position with respect to the laser robot coordinate system 19, that is, the tip position of the processing nozzle 15. Then, the homogeneous transformation matrix Π ₁ of the processing point to be obtained is as follows.

[0023]

(Equation 3)

The homogenous transformation matrices Π ₀ and Π ₁ are decomposed into movement amounts of the respective axes of the laser robot 14 and output as movement commands.

In the driving means 6, the processing nozzle 15
The servo motors of the respective axes of the laser robot 14 are driven based on the movement command output based on the equation (4) so that the distance between the workpiece robot 1 and the workpiece 1 becomes the preset distance L.

The servo motor of each axis of the laser robot 14 is driven by a movement command from the driving means 6, and the remaining distance Δd = 0 by the remaining distance calculation means 4, and the distance between the tip of the processing nozzle 15 and the processing work 1 When 1 becomes equal to the preset distance L, the processing proceeds to the processing after the shape sampling means 7. The shape sampling means 7, to sample the shape of the workpiece 1, as shown in FIG. 3, the processing nozzle 15 is rotated moving the control point O _C which is set in advance as a rotational center, first selected direction A Is sampled and stored as a shape data group 20 in the A direction.
Subsequently, sampling movement is performed in the B direction orthogonal to the A direction, and stored as a shape data group 21 in the B direction. As shown in FIG. 4, the processing movement of the control point to the control point O _C of the nozzle 15, change processing nozzle length of homogeneous transformation matrix [pi ₁ of the previously calculated expression (4) from N ₀ to N ₁ and, by newly calculating a homogeneous transformation matrix [pi _2, it can be easily realized.

[0027]

(Equation 4)

The selection of the center axis for rotational movement is as follows.
This is performed based on the inner product of each coordinate axis of the attitude of the processing nozzle 15 and the laser robot coordinate system 19. Selection axes ≡MIN selecting (vector Y _N · vector X _R, vector Y _N · vectors Y _R, vector Y _N · vector Z _R) or · · · (6) (X _R-axis or Y _R-axis or Z _R axis .)
The homogeneous transformation matrices Π ₃ , Π ₄ of the A-direction sampling start point 15 a and the A-direction sampling end point 15 b of the processing nozzle 15 shown in FIG.
_{Assuming that the R} axis and the preset rotation angle are θ _P , equation (5)
The following equation is obtained.

[0029]

(Equation 5)

(Equation 6)

As a result of the equations (7) and (8), each joint axis of the laser robot 14 is driven, and the processing nozzle 15 shown in FIG. 3 is moved from the A-direction sampling start point 15a to the A-direction sampling end point 15b. Perform a sampling operation.

The data sampled at a preset time interval is immediately transmitted to the processing nozzle 15 and the processing work 1.
And the control point O _{C of the} processing nozzle 15
Is calculated as a group of shape data of the processed workpiece 1 and stored in an auxiliary storage device (not shown). For example, assuming that the distance sampled at the sampling start point 15a in the A direction shown in FIG. 3 is h, the shape data S _A0 in the laser robot coordinate system 19 is as follows.

[0032]

(Equation 7)

By the same processing, the data group in the A direction is
Stored as _{A0 to} S _An . Equation (6) also applies to the B direction.
, The second smallest rotation axis is selected, the same processing is performed, and stored as the B-direction sampling data groups S _{B0 to} S _Bn .

In the next plane curved surface determining means 8, the direction A, the direction B
For each data group in each direction, a perpendicular line is drawn from each sampled data with respect to a straight line connecting the sampling start point and the end point, and the maximum value of the length of the perpendicular line is larger than a criterion ε preset by experience. Is a curved surface,
If smaller, it is determined as a plane. That is, as shown in FIG. 5, S _A0 ,
_Assuming that S _A1 ,..., S _An , the maximum perpendicular length in the A direction is
It becomes like the following formula.

[0035]

(Equation 8)

The same processing is performed for the B direction, and the maximum value of the perpendicular length in both the A and B directions is compared with the criterion ε.

When the plane is determined to be a plane by the plane curved surface determining means 8, the normal vector group calculating means 9 calculates four representative points S _A0 , S _An , S _B0 , and S _Bn from both A / B directions. Further, a group of normal vectors is calculated from data of a total of five points including the vector P ₁ as position data of the homogeneous transformation matrix Π ₁ of the processing point obtained by the equation (4). For example, when three points S _A0 , S _An , and S _B0 are selected from five points, the normal vector n ₁ is expressed by the following equation. Vector n ₁ = vector S _An S _A0 × vector S _B0 S _A0 ... (11) A combination of the processing of equation (11) that selects three points from five points ₅
C ₃ = 10 times, 10 normal vector group vectors n
Calculating a _1-vector n _10.

The next normal vector averaging means 10 calculates the average vector n ₀ of the vectors n ₁ to n ₁₀ of the normal vector group by the following equation to obtain the normal vector of the processing point on the plane.

[0039]

(Equation 9)

On the other hand, when the plane curved surface determining means 8 determines that the surface is a curved surface, the tangent calculating means 11 converts the machining point vector P ₁ shown in the equation (4) into the shape sampling means 7.
In each of the A / B directions, the tangent vector is calculated from the data of two points, respectively, in which of the data of the shape data group sampled in step (1). For example, the data group A direction as shown in FIG. 5, when it is sampled as a central axis X _R-axis, X _R component of the sampled data set are all the same value is entered. Therefore, in FIG. 5, it is sufficient to find out which data in the A direction data group surrounds the Y _R component of the vector P _{1 at} the position of the processing point. Here, _assuming that S _A8 and S _A9 are obtained, the tangent vector t ₁ is calculated as in the following equation. Vector t ₁ = Vector S _A8 S _A9 = (0, m ₁ , n ₁ ) (13) In the B direction, the tangent vector t ₂ is calculated by the same processing. Vector t ₂ = (l ₂ , 0, n ₂ ) (14)

The next normal vector calculating means 12 obtains the normal vector n ₁ and vector n ₂ from the tangent vector t ₁ and vector t ₂ , respectively.
Averaging is performed, and finally a vector n ₀ of an accurate normal vector to the curved surface is calculated. That is, Expression (13) and Expression (1)
4) From the following equation: Vector n ₁ = (0, n ₁ , −m ₁ ) (15) Vector n ₂ = (n ₂ , 0, −l ₂ ) (16) Therefore, the normal vector n _{0 of the} processing point Becomes as follows.

[0042]

(Equation 10)

After the processing in the normal vector averaging calculation means 10 or the normal vector calculation means 12, the position / posture registration means 13 calculates the position vector P ₁ of the processing point in the equation (4) and the plane surface. The expression (1
2) or the normal vector n ₀ in equation (17) is automatically registered as the position and orientation of the teaching data.

As described above, according to the automatic teaching device for a laser robot according to the present embodiment, even if the workpiece 1 is a flat or curved surface, the processing nozzle 15 of the laser robot 14 can be used.
Only by roughly positioning the position, the teaching along the normal vector of the processed workpiece, which is an important condition for laser processing, can be realized extremely accurately and easily.

<Embodiment 2> Next, an automatic teaching apparatus for a laser robot according to a second embodiment of the present invention will be described with reference to FIG. This embodiment corresponds to the embodiment of claim 2, and the same reference numerals and symbols are given to those having the same configuration or corresponding parts as those of the above-mentioned first embodiment, and detailed description thereof will be omitted. . As shown in FIG. 6, the shape sampling means 7 of the present embodiment performs sampling by moving in parallel along the coordinate axis of the laser robot coordinate system 19 selected by the equation (6). For example, as a result of equation (6),
When the X _R axis is selected in the A direction, the homogeneous transformation matrix at the A direction sampling start point 15c is Π ₅ , the homogeneous transformation matrix at the A direction sampling end point 15d is Π ₆ , and the processing point Assuming that a homogeneous transformation matrix is Π ₁ , and a predetermined parallel movement amount is D _P , the following equation is obtained.

[0046]

[Equation 11]

(Equation 12)

According to the equations (18) and (19), the sampling operation shown in FIG. 6 is performed, and the data sampled and moved in the A direction among the data sampled every moment is set as the shape data group 20 in the A direction. The data sampled and moved in the B direction orthogonal to the A direction is stored in an auxiliary storage device (not shown) as a shape data group 21 in the B direction.
The subsequent processing up to registration of the position / posture is the same as in the first embodiment.

As described above, according to the automatic teaching device for a laser robot according to the present embodiment, teaching along a normal vector of a processed workpiece can be realized extremely accurately and easily even for a processed workpiece having a small curvature. .

Embodiment 3 Next, an automatic teaching device for a laser robot according to a third embodiment of the present invention will be described with reference to FIG. This embodiment corresponds to the embodiment of claim 3, and the same reference numerals and symbols are given to those having the same configuration or corresponding parts as those of the above-mentioned first embodiment, and detailed description thereof will be omitted. . As shown in FIG. 7, in the shape sampling means 7 of this embodiment, for example, a contact type distance sensor 22 such as a stylus type sensor is attached to the tip of the processing nozzle 15, and the laser robot coordinate system selected by the equation (6). Sampling is performed by repeating parallel movement and movement in the direction of the processing nozzle along the 19 coordinate axes.
For example, the result of the expression (6), X _R-axis as a direction A, the D _P parallel movement amount and approach height H _P into workpiece 1, a homogeneous transformation matrix of the working point from equation (4) Π
_If it is set to ₁ , first, the homogeneous transformation matrix Π ₇ of the approach start point 15 e of the machining nozzle 15 for approaching the workpiece 1 is calculated as in the following equation.

[0050]

(Equation 13)

Then, from ら れ₇ obtained by equation (20), approach is made in the direction of the processing nozzle, and position data when a contact signal is input from the contact type distance sensor 22 is used as shape data. That is, assuming that the moving amount at the time of approaching is A (t) as an amount that changes with time, the homogeneous transformation matrix ア プ ロ ー チ₈ during the approaching is as follows.

[0052]

[Equation 14]

As soon as the contact signal is input, the position data vector P ₈ is stored as shape data S _A0 in an auxiliary storage device (not shown). As described above, the sampling operation of repeating the parallel movement and the movement in the processing nozzle direction of Expressions (20) and (21) is performed in each of the A direction and the B direction. A data group 20 is obtained by sampling and moving in the B direction orthogonal to the A direction, and the shape data group 21 in the B direction
Stored as Subsequent processing up to registration of the position / posture is the same as in the above-described first embodiment.

As described above, according to the automatic teaching device of the laser robot according to the present embodiment, even if the non-contact type distance sensor cannot be used for a work such as wood, the normal vector of the work is used. Can be realized very accurately and easily.

[0055]

As described above, in the automatic teaching device for a laser robot according to the first aspect, after the laser robot is roughly positioned, the remaining distance calculating means and the processing nozzle direction calculating means process the laser robot with respect to the processing point. After accurately determining the position of the nozzle, it is equipped with shape sampling means, plane curved surface determination means, and means for calculating normal vectors corresponding to planes and curved surfaces, and teaching data for machining points is registered very accurately and easily. can do.

In the automatic teaching device for a laser robot according to the second aspect, in addition to the effect of the first aspect, the shape sampling means further performs sampling on a work piece having a small curvature by parallel movement along the laser robot coordinate system. However, the teaching data for the processing point can be registered very accurately and easily.

According to a third aspect of the present invention, in addition to the effect of the first aspect, the shape sampling means further includes a contact type distance sensor for performing parallel movement along the laser robot coordinate system and processing nozzles. By sampling by a combination of movements in the directions, teaching data can be registered very accurately and easily even for a work piece made of a material that cannot be sampled by a non-contact distance sensor.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an entire configuration of an automatic teaching device for a laser robot according to a first embodiment of the present invention.

FIG. 2 is an explanatory diagram showing a relationship between a laser robot coordinate system and a processing nozzle coordinate system in the automatic teaching device for a laser robot according to the first embodiment of the present invention.

FIG. 3 is an explanatory diagram showing an operation of a processing nozzle in the automatic teaching device of the laser robot according to the first embodiment of the present invention.

Figure 4 is an explanatory diagram showing a control point O _C of the processing nozzle in the automatic teaching apparatus according laser robot to the first embodiment of the present invention.

FIG. 5 is an explanatory diagram showing a normal vector n ₀ of a processing point in the automatic teaching device for a laser robot according to the first embodiment of the present invention.

FIG. 6 is an explanatory diagram showing an operation of a processing nozzle in the automatic teaching device of the laser robot according to the second embodiment of the present invention.

FIG. 7 is an explanatory view showing the operation of a processing nozzle in the automatic teaching device for a laser robot according to the third embodiment of the present invention.

FIG. 8 is an explanatory diagram showing a concept in teaching of a conventional laser robot.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Processing work 2 Detection data conversion means 3 Distance calculation means 4 Remaining distance calculation means 5 Processing nozzle direction calculation means 6 Driving means 7 Shape sampling means 8 Plane curved surface judgment means 9 Normal vector group calculation means 10 Normal vector averaging calculation means Reference Signs List 11 tangent calculation means 12 normal vector calculation means 13 position / posture registration means 14 laser robot 15 processing nozzle 16 visible light 17 (non-contact type) distance sensor 18 processing nozzle coordinate system 19 laser robot coordinate system

Claims (3)

(57) [Claims] 1. An automatic teaching device for a laser robot capable of moving a processing head for irradiating a laser beam guided from a laser oscillator onto a processing work having an arbitrary plane or a curved surface in an arbitrary direction. One non-contact distance sensor attached to the processing head facing a processing point; detection data conversion means for converting analog data from the distance sensor into digital data; and the digital data from the detection data conversion means. Distance calculating means for calculating a distance between the tip of the processing head and the workpiece based on data; calculating a difference between the distance calculated by the distance calculating means and a distance required for a predetermined proper processing; Remaining distance calculation means, and moving the processing nozzle closer to the processing work by the difference calculated by the remaining distance calculation means. Processing nozzle direction calculating means for calculating a movement command to be performed; driving means for driving a servo motor of the laser robot based on the movement command calculated by the processing nozzle direction calculating means; After the processing head reaches a distance based on the shape data, the processing head is rotated and moved at a preset control point, and a shape sampling unit that calculates a shape data group; and the shape data group calculated by the shape sampling unit. A plane curved surface determining unit that determines whether the surface is a plane or a curved surface; and when the plane curved surface determining unit determines that the surface is a plane, five representative points are selected from the shape data group sampled by the shape sampling unit. A normal vector group calculating means for calculating a normal vector group for the processed workpiece; and a normal vector group calculating means. An average vector of the calculated normal vector group is calculated, and a normal vector averaging calculation unit as a normal vector of the processing point, and when the plane surface determination unit determines that the surface is a curved surface, the shape is determined. A tangent calculating means for calculating a tangent from the shape data group sampled by the sampling means and the processing point; and a normal vector calculation for calculating a normal vector of the processing point from the tangent calculated by the tangent calculating means. Means, and a position corresponding to the distance set by the driving means is a position of the laser robot at the processing point, and the normal vector calculated according to whether the shape data group is a plane or a curved surface is processed by the processing. A position / posture registration unit for registering the position of the laser robot at a point. Automatic teaching apparatus of the bot. 2. The apparatus according to claim 1, wherein said shape sampling means moves the machining head in parallel with a predetermined width along a designated coordinate axis of a laser robot coordinate system to calculate a group of shape data. 2. The automatic teaching device for a laser robot according to 1. 3. The method according to claim 1, wherein the distance sensor is a contact type, and the shape sampling means is configured to move the processing head, to which the distance sensor is attached, by a predetermined width along a specified coordinate axis of a laser robot coordinate system. Move in combination with the movement of the preset distance in the direction of the processing nozzle,
2. The automatic teaching device for a laser robot according to claim 1, wherein a shape data group is calculated.