JPH0123280B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a playback type industrial robot, and particularly to one that facilitates teaching.

Industrial robots that use a playback method to automatically control the position and posture of their end effectors are well known.

However, in this teaching, the position and posture of the end effector are controlled by manually operating a remote control switch, for example, and this information is stored in an attached computer. However, in this case, accurate manual control of the end effector is not only difficult, but also requires considerable skill and time, resulting in poor work efficiency. Furthermore, as an improvement on this, a method has been considered in which the end effector can be directly held and moved by the operator, but this also requires each control axis to be free, and there are various difficulties involved in realizing it, so it cannot be put into practical use. It has not been.

Therefore, the inventor of this invention proposed that the end effector is not directly held by the operator, but a corresponding model is prepared separately, the operator holds it in a desired position and posture, and while the end effector or Either one of the replaced dummies is equipped with means for recognizing the mutual positional relationship between multiple points on the model and the end effector or the dummy, and by inputting the output of this recognition means into a computer, This invention was completed based on the idea that the above-mentioned problems could be solved if the computer was made capable of calculating the position and orientation of the model.

A first embodiment of the present invention will be described in detail below with reference to FIG.

1 is a well-known rectangular coordinate (X _R ,
Y _R , Z _R ) are the vertical axes configured at the terminals of the robot R.

A first arm 2 is supported at the lower end of the vertical shaft 1 so as to be able to pivot α around the shaft 1.

A second arm 3 is supported at the tip of the arm 2 by an oblique shaft 3a so as to be rotatable β. At the tip of the second arm 3 is an end effector (welding torch T in this example).
A gripping tool 3b for gripping is provided.

The shafts 1, 3a, and the central axes of the torch T are configured to intersect at one point P. Furthermore, the torch T is configured such that its welding operating point can coincide with the point P. Thus, by controlling the angles α and β, the attitude angle θ and the turning angle φ of the torch T with respect to the vertical axis 1 can be controlled.

4 is a model. Model 4 is state 4
1. The collar 42 and the grip 43 are integrated. The tip 41a of the stake 41 is configured as a light emitting portion. Further, the center of the disc of the collar 42 is fixed at a right angle to the stake 41, and the periphery thereof is configured as a circular light emitting portion 42a. The grip 43 is fixed at the proximal end portion of the stake 41 opposite to the distal end 41a. 43a is grip 4
This is the start switch provided at 3. Although the details of the light emitting parts 41a and 42a are not shown, they guide light from a light source in the grip 43 using an optical fiber. In particular, the light emitting part 41a has a light-opening surface at the end of the optical fiber, and the light emitting part 42a The optical fiber is shaped like a ring, and the surface of the optical fiber is made up of a linear satin-finished part.

Note that the diameter of the light emitting part 42a is D, and the length from the collar 42 of the stake 41 to the tip 41a is h.
Furthermore, from the collar 42 of the stake 41 to the base end 41b
Let the length up to that point be h ₁ .

CO is a computer attached to this industrial robot R, and includes a CPU and memory. A servo system SX for the X _{and R} axes of the robot R is connected to the bus line B of the computer CO. The servo system SX includes the power MX for the X _{and R} axes, and the encoder EX that outputs its position information. Similarly, the bus line B is connected to a similarly configured Y _R- axis servo system SY, Z _R -axis servo system SZ, α-axis servo system Sα, and β-axis servo Sβ. RE is a remote control panel and a group of manually operated snap switches.
Install SW. If you turn the snap switch for each control axis of X _R , Y _R , and Z _R to the "U" side, the position information for that control axis will increase, and if you turn it to the "D" side, the end effector will move in the opposite direction. Configured to move. Further, the snap switches corresponding to the α and β control axes are configured such that when pushed to the "C" side, the end effector is rotated clockwise, and when pushed to the "CC" side, the end effector is rotated counterclockwise.

The operation panel RE is also provided with a speed command rotary switch SV. In addition, a mode changeover switch SM is provided, allowing manual mode MA and model mode MO.
and is configured to be able to switch to auto mode A.

Further, it is assumed that the start switch 43a and the switches on the operation panel RE described above are connected to the computer CO via the bus line B.

TD is a dummy that can be installed in place of torch T. This dummy TD is configured as a solid-state industrial television camera. It is desirable that the camera be a fixed focus camera. Further, its optical axis is made to coincide with the central axis of the torch T. DF is the image sensor of this TV camera TD. This television camera also has fixed orthogonal coordinate axes X, Y, and Z, the origin of which is the principal point of the lens, and one of the coordinate axes is parallel to the direction of the element matrix of the image sensor DF. . Furthermore, let the focal length of this television camera be γ.

Also, on the surface of the image sensor DF,
The two-dimensional orthogonal coordinate axes of y are fixed, and the x-axis and y-axis are set in opposite directions to the X-axis and the Y-axis, respectively.

The operation of the above-mentioned embodiment will be explained below.

First, a dummy TD is attached to the gripper 3b instead of the torch T (as shown by the solid line in FIG. 1).

Then, set the switch SM on the operation panel RE to manual mode MA, operate the switch group SW on the operation panel RE, manually operate the position and orientation of the dummy TD, and weld the weld line of the workpiece W to be taught.
Close to welding point on WL.

And set Switch SM to model mode MO,
The operator places model 4 at its proximal end 4 as shown.
1b is brought into contact with the welding point to be taught, and the direction of the stick 41 is set as the attitude that the torch T should take.

In this case, the images of model 4 on image sensor DF are as shown in FIG. 2, 41aF and 42aF.

Then, the operator operates the switch 43a. Computer CO receives this start signal and
The position and orientation of the model 4 are calculated from the images 41aF and 42aF. That is, the coordinates of the image 41aF on the image sensor DF are (xm, ym), the coordinates of the center of the image 42aF on the image sensor DF are (xw, yw), and the coordinate system of the light emitting part 41a (X, Y, Z ), the coordinate values in the coordinate system (X, _Y , _Z ) of the _center of the light section 42a are (X _W , Y _W , Z _W ), the image 42 aF If the length of the major axis of the ellipse is d, then d/D=γ/Z _W =1/A where Z _W =A〓 X _W =Axw Y _W =Ayw X _M =Z _M /γxm Y _M =Z _M /γym (X _W −X _M ) ² + (Y _W − Y _M ) ² + (Z _W − Z _M ) ² = h ² , Here, a=1+1/γ ² (xm ² +ym ² ) b=Aγ+A/γxwxm+A/γywym c=A ² (xw ² +yw ² +γ ² )−h ² Therefore, the coordinate value of point 41b is 〓=(〓 −〓) h ₁ /h+〓 Also, the attitude vector of the stake 41 becomes: 〓=〓−〓/｜〓−〓｜.

Furthermore, these coordinate values 〓 and 〓 are coordinate-transformed into values 〓' and 〓' of the robot R's absolute coordinate system (X _R , Y _R , Z _R ). That is, 〓′=〓 _T 〓 〓′=〓 _T 〓 However, 〓 _T is a known coordinate transformation matrix with 4 rows and 4 columns, and the value of each element is the length of arms 2 and 3, and the values of α and β. This is the value determined by The details of this coordinate transformation will not be described in detail because it is conventionally well known.

In the same manner, information on the desired teaching point on the welding line WL can be taught by performing the same operation.

In this embodiment, the dummy TD can also be made to automatically follow the movement of the model 4.

That is, as mentioned above, the computer CO
calculates the position information of the point 41b of the model 4 and the posture information of the stake 41, so the welding position P on the center line of the dummy TD is overlapped with the point 41b, and the posture of the center line of the dummy TD is made to match that of the stake 41. , each servo system SX, SY, SZ, Sα
and Sβ, thus
It will be appreciated that the dummy TD can be made to follow the movement of Model 4. In this case, each time the operator operates the switch 43a, the teaching operation can be executed by storing the information for each control axis at that time in the memory of the computer CO.

Next, as another example, model 4 and dummy
Another embodiment of the TD will be described mainly regarding differences from the previous embodiment with reference to FIGS. 3 to 5.

In this example, model 4 has its stand 4
1, known ultrasonic wave transmitting elements TX ₁ and TX ₂ are provided. These elements TX ₁ and TX ₂ are powered by a battery in the grip 43 or by a cord from a source (not shown) passing through the grip 3.

It is desirable that the dimensions of elements TX ₁ and TX ₂ be as small as possible, and the spacing between them is approximately 20 mm. Also, the oscillation frequency is the same, approximately 0.5 to 1M
Approximately Hz. Note that the distance between the elements TX ₁ and TX ₂ is h, and the distance between the element TX ₂ and the base end 41b is h ₁ .

On the other hand, known wave receiving elements R ₁ , R ₂ , R ₃ and R ₄ capable of receiving the sound waves of elements TX ₁ and TX ₂ are attached to the dummy TD, arranged at each vertex of an equilateral triangular pyramid.
It is desirable that the dimensions of each of these elements R ₁ , R ₂ , R ₃ and R ₄ be as small as possible. Also, the distance between them is approximately 50 mm. Although not shown, each element R ₁ to R ₄ is supported by a thin steel wire from a dummy TD. The output information of each of these elements _R1 to _R4 is connected to the byline B as in the previous embodiment.

Furthermore, the supply voltage to elements TX ₁ and TX ₂ is determined by the delay time TL as shown in FIG.
shall be performed by a rectangular wave voltage having
As shown in the figure, the output waveform is a rectangular wave modulated by the vibration waveform of the element. Time TL is about 50
~100msec is desirable.

To describe the operation of this embodiment described above, each element R ₁ to R ₄ receives the oscillated sound waves from elements TX ₁ and TX ₂ , and based on the reception time interval, each element R 1 to _{R 4} receives the oscillated sound waves from elements TX 1 and TX 2. R ₄ to each element TX ₁ and
The fact that the distance to TX ₂ can be determined from information input into a computer is well known and will not be described in detail.

Then, L = distance between each element R ₁ to R ₄ li ₁ = measured distance between element TXi and R ₁ (i = 1, 2) li ₂ = measured distance between element TXi and R ₂ (i = 1, 2) li ₃ = Measured distance between element TXi and R ₃ (i = 1, 2) li ₄ = Measured distance between element TXi and R ₄ (i = 1, 2) 〓〓 ₁ = ₁ point of element TX local coordinate system (dummy
position vector in a coordinate system fixed to TD).

〓〓 ₂ = Position vector in local coordinate system indicating _two points of element TX.

〓 ₁ = Position vector in the local coordinate system indicating _one point of element R.

〓 ₂ = Position vector in local coordinate system indicating _two points of element R.

〓 ₃ = Position vector in local coordinate system indicating ₃ points of element R.

〓 ₄ = Position vector in local coordinate system indicating ₄ points of element R.

If we further code as shown in Fig. 5 a and b, the position vector of the teaching point 41b is 〓=(〓〓 ₂ −〓〓 ₁ )h ₁ /h+〓〓 _2Also , the posture of the teaching point 41 is The vector 〓 shown is 〓=(〓〓 ₂ −〓〓 ₁ )/｜〓〓 ₂ −〓〓 ₁ ｜ 〓〓i (i=1, 2) is derived as follows.

b=0.5L δ ₃ =√ ² − ² ai=li ² / ₃ +L ² −li ² / ₂ /2L di=ai−b=li ² / ₃ −li ² / ₂ /2L σi=√ ² ₃ − ² σi=√ ² ₁ − ² ci=σi ² / ₁ +σ ² / ₃ −σi ² / ₂ /2σ ³ σi ₄ =√ ² ₁ − ² 〓=b〓 ₃ −〓 ₂ /〓 ₃ −〓 ₂ +〓 ₁ −〓 ₃ 〓=〓／｜〓｜ 〓i＝〓 ₃ −ai〓 ₃ −〓 ₂ /｜〓 ₃ −〓 ₂ ｜＋〓ci 〓=〓 ₄ −〓 ₁ −〓 ₂ +〓 ₃ /3 〓= 〓／｜〓｜ Ai ₁ =｜〓｜+σi ₄ Ai ₂ =｜〓｜−σi ₄ 〓i＝〓 ₁ −〓 ₂ +〓 ₃ /3−〓i γi ₁ =√｜〓｜ ² + ² ₁ γi ₂ ^{= √} _{| _ _ _ _ _} 〓i=〓(When |li ₄ −γi ₁ | > |li ₄ −γi ₂ |) Thus, in this embodiment as well, as in the previous embodiment, the operator holds the model 4 at an arbitrary position and orientation. However, the information at this time can be stored in the memory of the computer CO.

In addition to arranging the elements R ₁ to R ₄ at each vertex of the equilateral triangular pyramid, they may be arranged at any four locations. In this case, the above formulas become somewhat complicated.

Next, as a third example, model 4 and a dummy
Another embodiment of the TD will be described mainly regarding differences from the previous embodiment with reference to FIG.

In the model 4, rods R ₁ and R ₂ are supported at two positions by ball joints J ₁ and J ₂ so as to be rotatable in three dimensions.

And for the dummy TD, these rods R ₁ and R ₂ are set to 3
Supports S ₁ and S ₂ are provided which bear on the dimensions.

The receivers S ₁ and S ₂ are bearings SJ ₁ and SJ ₂ that can slide the rods R ₁ and R ₂ , and a universal joint U ₁ that can rotate these bearings SJ ₁ and SJ ₂ in three dimensions.
and U ₂ , consisting of.

Furthermore, both universal joints U ₁ and U ₂ are supported by the dummy TD in the axial direction of the dummy TD and in a direction perpendicular thereto. In addition, bearings SJ ₁ and SJ ₂ have a length of protrusion of rods R ₁ and R ₂ (from joint J ₁ or J ₂ to the intersection of the orthogonal axes of bearing SJ ₁ or SJ ₂ , that is, the rotation fulcrum). The sensor is equipped with a detector (not shown, but a known means such as a linear encoder) capable of detecting the distances l ₁ and l ₂ ). Further, the joints U ₁ and U ₂ have rotation angles ψ ₁ and θ ₁ as shown in the figure,
and a detector (not shown, but a known means such as a shaft encoder) capable of detecting the rotation angles ψ ₂ and θ ₂ .

The output information from these detectors is then connected to the computer CO via bus line B.

Further, a rectangular coordinate system x, y, z unique to the dummy TD is fixed as shown in the figure. and z
The axes are assumed to be in the same direction as the ψ ₁ and ψ ₂ axes, and the y axis is in the same direction as the θ ₁ and θ ₂ axes.

In this embodiment, the position vector T of the teaching point 41b of the stake 41 is: 〓=(〓 ₂ − 〓 ₁ )h ₁ /h+〓 ₂ , and its posture vector is: 〓=〓 ₂ −〓 ₁ /｜〓 ₂ − 〓 ₁ | Here, 〓i=〓i+xi〓+yi〓+Zi〓(i=1,2) Also, xi=l icosθi cosψi yi=l isinψi zi=l icosψi sinθi Furthermore, 〓, 〓, 〓 are respectively x, Denote the unit vector of the y and z axes 〓i (i.e. 〓 ₁ and 〓 ₂ )
represents the positional information regarding the positions of the rotational fulcrums of universal joints SJ ₁ and SJ ₂ regarding the x, y, and z axes.

Thus, in this embodiment as well, if the operator brings the dummy TD near the teaching point by manual operation, holds the model 4 at the desired teaching position, and operates the switch 43a, the above-mentioned desired teaching position can be reached. Posture information is stored in the memory of the computer CO.

In the above embodiment, the number of rods is two, but it is also possible to use more, for example, three rods, and the corresponding three universal joints on the dummy side are provided with means for detecting the protrusion length of each rod. , the position and orientation of the model can be recognized. Furthermore, as a modification of these methods, a string or the like may be stretched between the model and the dummy, the dummy side may wrap the string, and the length of the stretched string or the like may be detected.

The present invention is not limited to the embodiment described above; for example, in the embodiment described above, a dummy model recognition means is provided, but this is because it is often difficult to provide it directly to the end effector. however,
If possible, the end effector may be provided with direct recognition means.

Replacement of each component with an equivalent within the scope of the technical idea of this invention is also included within the technical scope of this invention.

Since this invention is as described above, an operator can hold a model other than an industrial robot in a desired position and posture and utilize the excellent operability of the human hand by simply pressing a switch. Therefore, the playback method of teaching can be easily executed and can produce unique and remarkable effects.

[Brief explanation of drawings]

1 and 2 show a first embodiment of the present invention, with FIG. 1 being a side view of the main parts including a block diagram, and FIG. 2 being an explanatory view of the action seen from the arrows in FIG. 1. 3 to 5 a and b show the second embodiment, in which FIG. 3 is a side view of the main part, and FIGS. 4 and 5 a and b are explanatory diagrams of the operation. FIG. 6 shows a third embodiment, and is a side view of the main part. CO...Computer, T...Torch (end effector), R...Industrial robot, 4...Model, TD...Dummy, 41a and 42a...
Light emitting part, DF...Image sensor, TX ₁ and
TX ₂ ... Ultrasonic wave transmitting element, R ₁ , R ₂ , R ₃ and R ₄
... Receiving element, R ₁ and R ₂ ... Rod, S ₁ and S ₂ ... Receptacles that support the rod in three dimensions.

Claims (1)

[Claims] 1. In a playback type industrial robot that is attached with a computer and teaches the position and orientation of the end effector in advance, a model corresponding to the end effector is prepared separately, and Either one of the replaceable dummies is provided with means for recognizing the mutual positional relationship between a plurality of points on the model and the end effector or the dummy, and the computer inputs the output of the recognition means. The industrial robot is configured to calculate information on the position and orientation of the model. 2. The industrial robot according to claim 1, wherein the model includes a plurality of light emitting sections, and the end effector or dummy further includes an image sensor that recognizes the light emitting sections. 3. The model includes a plurality of ultrasonic wave transmitting elements, and the end effector or dummy further includes a plurality of wave receiving elements capable of receiving and recognizing ultrasonic waves from the ultrasonic wave transmitting elements. An industrial robot according to claim 1. 4 The model has one end of a plurality of rods supported by a ball joint, and the end effector or dummy has a receiving member that slidably supports the other ends of the rods by a universal joint. The industrial robot according to claim 1, further comprising means for detecting at least the length of the rod.