JPS61269709A - Operation control system for automatic operating device - Google Patents

Abstract

PURPOSE:To decrease taught coordinate positions even when microinstructions are used by providing a state detecting means which detects the state of a working part, a memory stored with working instructions and coordinate values, and a control part which performs movement control according to the contents of the memory. CONSTITUTION:When a magnetic disk plate 9 is taken out, a working instruction MC indicates two steps and a control part CT controls the movement of a moving means MT from a point A to a point B with a movement command C using indication coordinates of the memory 81 to position a hand 5 at an indicated coordinate position B, then controls the movement of the moving means MT with a movement command V to move the hand 5 toward an object member 9, and receives a state quantity F fed back from a state sensor 3 to position the hand 5 at a position C according to the state of the member 9. Thus, the two steps are executed on one-position coordinates with one instruction. Similarly working operation for sucking the plate 9 by the hand 5 and the elevating and positioning operation for the hand 5 sucking the plate 9 at the point B are carried out with one instruction. The position of the point B at this time is already given with the last instruction, so it need not be given.

Description

[Detailed description of the invention] [Table of contents] Overview Industrial field of application Conventional technology Problems to be solved by the invention Means for solving the problems (Fig. 1) Working example (α) One example Explanation of the configuration (Figures 2 and 3) (A)
Explanation of the configuration of the control unit (Figure 4) <c> Explanation of teaching data (Figure 5) (d) Explanation of the operation of the configuration of one embodiment (Figures 6, 7, 8, and 9) ) (e) Explanation of another embodiment (Fig. 10) (a) Explanation of another embodiment Effects of the invention [Summary] In an automatic working device that works on a target member by moving a working part within a working space By providing a state detection means for detecting the state of the working part with respect to the target member, a memory for storing work commands and coordinate values, and a control section for controlling movement based on the contents of the memory, the coordinate value can be detected with one work command. 1st to
, and a subsequent movement movement in the direction of the target member.

[Industrial application field]

The present invention relates to an operation control method for controlling the movement of a moving means according to the teaching contents of a memory in an automatic working device that performs work by moving a working part of a robot or the like using a moving means.
In particular, the present invention relates to an operation control method for an automatic working device that enables a series of multiple moving operations even when the teaching content is one work command and one coordinate value.

Object moving devices that move part of the work are widely used, and a typical example is a robot.
A hand serving as a working part is positioned at a desired position, thereby making it possible to perform desired work.

In this type of robot, the robot's memory bag can be created by teaching the movement commands and position/attitude data necessary for tasks such as the position the hand should pass, the posture of the hand, and the movement. On the other hand, during playback, the movement commands and position-posture data are sequentially read out from the storage device and the arm and hand are controlled to reproduce the movement as taught. There is. This type of robot control method is called a teaching playback method.

[Conventional technology]

For example, in the work of removing a magnetic disk plate in the assembly work of the magnetic disk device shown in FIG. Sequence operation is required.

(1) Move to 0 point and position. (MOVE R)
(II) Move to point O and position. (MOVW
C) (iiD Vacuum adsorption o (VA
CUUM 0N)Ov) Move to 0 point and position. (MOVE R) In other words, four movement commands and two coordinate values t were required to perform one extraction operation. In addition, when a plurality of magnetic disk plates 9 are housed in the jig 6, the number of magnetic disk plates 9 needed to be set as the zero point is necessary, which is a problem.

In order to save this kind of teaching effort, a robot language (T/necessity) is required that uses macro commands as work commands.

For example, the magazine ``Nikkei Mechanical 1985, 4゜8 No. J (
Published by Nikkei McGraw-Hill, Inc., April 8, 1985, No. 7
As seen in the literature on robots from pages 3 to 81, the need for a robot language that uses macro instructions for describing multiple actions is shown in order to make teaching programming easier.

If we introduce the idea of such macro instructions into the sequence operation described above, for example, steps (1) and (II) can be
Step (iiDO) is written in one macro instruction.

[Problems to be Solved by the Invention] However, simply converting a series of multiple operations into macro commands does not reduce the number of coordinate positions to be taught, but merely simplifies the work commands. There is a problem in that the time and effort required for teaching does not decrease.

SUMMARY OF THE INVENTION An object of the present invention is to provide an operation control method for an automatic working device that can reduce the number of taught coordinate positions even by using macro instructions.

[Means for solving problems]

FIG. 1 is a diagram explaining the principle of the present invention.

In FIG. 1 (8), 3 is a state detection means, which detects the state of the working part with respect to the target object, and is composed of, for example, a force sensor that detects an external force applied to the working part, and 5 is a state detecting means. This is a working unit, and is, for example, a hand that performs work on a target member. MT is a moving means that moves the working part (hereinafter referred to as hand) 5 within the work space, and is composed of a motor etc., 8 is a control device,
A control unit CT that controls the movement of the moving means MT, and controls the movement of the moving means according to the memory 81 that stores teaching contents and the state quantity from the teaching contents and the state detection means (referred to as a state sensor) 3. and 9 is a target member, which is a target for the hand 5 to work on.

The memory 81 contains work instructions MC written in macro instructions.
and the corresponding coordinate position B are stored, and this work command Me instructs at least a series of a plurality of moving operations. The control unit CT is configured to execute a series of multiple movement operations in response to the work command Me.

[Effect]

In the present invention, the control unit CT executes a series of a plurality of moving operations in response to the work command MC.

This series of multiple operations moves the hand 5 to the corresponding coordinate position.
and a second movement operation in which the hand 5 is moved from this coordinate position in the direction of the target member 9 at a constant speed while movement is controlled by the state quantity of the state sensor 3.

For example, in the work to take out the magnetic disk plate 9 shown in FIG.
1), (II), and the control unit CT moves from point A to point B.
The movement of the moving means MT is controlled by the movement command C using the specified coordinates of the memory 81 until the hand 5 is moved to the specified coordinate position tB.
Next, the control unit CT controls the movement of the moving means MT using the movement (fast-reverse) command V, moves the hand 5 in the direction of the target member 9 (downward), and performs feedback based on the state quantity of the state sensor 3. Then, the hand 5 is positioned at the 0 point based on the state (contact or proximity) with the target member 9.

Therefore, steps (+) and (II) described above can be executed with one command and one position coordinate.

Similarly, step (III) li Gφ can be done with one command, and in this case, the working operation of the hand 5 to attract the magnetic disk plate 9, and the third operation of raising and positioning the hand 5 which has attracted and held the magnetic disk plate 9 to the 0 point are performed. A moving action is performed.

Since the position of point B at this time has been given in the previous instruction, there is no need to provide it separately.

Furthermore, steps (1) to u(IV) can be executed with one instruction and -coordinate. For example, if this is an "rPIcK" command, the control unit CT performs the first moving operation of moving the hand 5 to point B, lowering the hand 56, :! c-181 - 4-movement operation of the housing 2 - A series of working operations to attract the magnetic disk plate 9 to the hand 5, and a third movement operation to return the hand 5 to point B while holding the magnetic disk plate 9. The operation sequence for extraction can be realized by one macro instruction and one coordinate.

〔Example〕

(α) Description of configuration of one embodiment FIG. 2 is a configuration diagram of one embodiment of the present invention, taking a Cartesian coordinate robot as an example.

In the figure, the same parts as shown in Fig. 1 are indicated by the same symbols, and ]α and 1h are X-axis modules, which constitute two X-axis positioning mechanisms of the robot, and each have an X-axis motor. 10α, 10h transport pallets] 1α, 11b are transported and positioned in the X-axis direction; 2 is a gate-type robot; a pair of support bases 20, 21 provided on both sides of the The two-axis block 22 that moves in the axial direction, the two-axis movable part (arm) 23 that moves in the X-axis direction, and the two-axis block 22 are fed and the ball screw 24 is moved.
A Y-axis motor 24 rotates α and drives in the X-axis direction along guides 25α and 25b, and a two-axis arm 23 provided on the two-axis block 22 is driven in the Z-axis direction via a ball screw feeding mechanism (not shown). It has a two-axis motor 26. Reference numeral 3 designates a force sensor as a state sensor, and as shown in FIG. 5 is a vacuum suction hand, which has a suction surface at the tip of the cylindrical body and has an intake tube connected to an intake pump. , 6 is a jig, which fixes the (magnetic disk plate) disc 9 on the pallet 11α, and 7 is a base, which is mounted on the pallet 11h and to which the disc 9 is attached.

Reference numeral 80 is an operation panel, which is operated by the operator to instruct playback mode, teaching mode, etc., 81 is a memory, which stores teaching contents, and is detailed in FIG. 5; A processor (hereinafter referred to as CPU) is composed of a microprocessor, etc., and reads out the contents of the memory 81 during playback and gives commands to each part. Reference numeral 83 is a servo control unit, and the X-axis module 1
α.

1b's X-axis motors 10α, 106 and Y-axis motor 24,
In order to control the position and speed of the z-axis motor 26 and the γ-axis motor 4, the command position cx2. ax,, ay
, cz , cγ are input, and the current position t PX2゜PXI , PY, PZ ,
Pr is fed back, and furthermore, the command speed V is sent from the synthesis section.
cm to vcr are input, and servo control is performed by these. 84 is a power amplifier, which amplifies the input and outputs
,Y,Z.

85 is a combining unit that supplies current to the γ-axis motors 10α, 10b, 24, and 26.4, and as described later in FIG. 87 is a force control unit that calculates the difference between the force control command PFX-PFγ of the control unit and supplies it to the servo control unit 83, and as described later in FIG. and converts it into a digital value F~F, sets a dead zone, and outputs a control output PFX-PFγ,
88 is a hand position detection unit, and motors 10α for each axis;
10b, 24, 26.4, the current position of each axis PX1. PX2, PY
, PZ, and PY to obtain the current position of the hand 5. 89 is a bus, which includes a CPU 82, a memory 81, an operation panel 80, a servo control section 83, a synthesis section 85, and a force control section 8.
7 and hand position detection circuit 88 to exchange data and commands.

FIG. 3 is a configuration diagram of the four-degree-of-freedom sensor 3 in the configuration shown in FIG.

The force sensor 3 detects external forces on the x, y, and z axes.
z force detection module 740 and r for detecting the external force on the γ axis.
and a force detection module 750.

As can be seen from FIG. 3, the force detection module 740
Since the displacement directions of each parallel plate spring body are perpendicular to each other, the deflection in the X-axis direction occurs in the parallel plate spring bodies α1 and α1′, and the deflection in the X-axis direction occurs in the parallel plate spring bodies b 1 and b 1′. It has three degrees of freedom in which the parallel plate spring body C1*C1' shares the deflection in the X-axis direction and the deflection in two directions, respectively.

Supports 743 and 744 respectively support the force detection module 740, and the support 743 is connected to the square rod 742 by a screw 745, and the support 744 is connected to the square rod 741 by a screw 746.

Note that only one side of the screws 745 and 746 is shown, and the screw hole 743α into which each screw 745 is screwed and the other hole are set at positions ζζ that are equal distances from the center position of the center hole 740α, and the screws 746 are similarly screwed together. screw holes 744α and 744
b is an equal distance from the center position of the center hole 250α (L9=
L10).

747 is the support body 743! The output rod is connected by a ζ screw 748 and is configured to pass through a hole 740α provided in the force detection module 740.

In this case, the support body 744 is fixed to the vacuum suction hand 5. Note that the output rod 747 is configured to pass through a hole 740α provided in the force detection module 740, or is configured to penetrate the support body 744. In this case,
The support body 744 is attached to the base on the opposite side (square bar 741.
742 side).

749α　, 749b, 749t'', 749d.

7496.749f is a strain gauge, and each parallel leaf spring body α1', h1. Detecting the displacement of CI'0Here, in order to detect the force in the axial direction without being affected by torque, this strain gauge is attached so that it is symmetrical with the center hole 740α as the center, and the bridge circuit is Therefore, although not shown, the parallel plate spring bodies α1, C1
, b1' are also attached with two strain gauges so as to correspond to the center point of the center hole 740α.

With the configuration described above, for example, the output rod 7
47 is a strain gauge 749C when a force in the X-axis direction is applied.
, 749d can detect the displacement of the parallel plate spring C'j and can detect force only in the X-axis direction, and similarly when a force in the Y-axis direction is applied, the strain gauge 749g.

749f detects the displacement of the parallel plate spring b1, and when a force in the z-axis direction is applied, the strain gauges 749α and 749b detect the displacement of the parallel plate spring α'1, and detect force components in each axis.

Furthermore, even if the resultant force from multiple directions is applied, the square bar 741
, 742 is applied by the supports 743° 744 at equal distances from the center position of the center hole 740α, so that each parallel plate spring body exerts its respective component force F x , F
750 is a γ force detection module that can independently detect y, and includes a central member 752 attached to the output rod 747 of the force detection module 740 via a screw 751, and a leaf spring 750α,
It includes an outer ring 753 connected via 750b, 750D, and 750d. 754α.

754b, 754C, and 754d are strain gauges, which are attached to the leaf springs 750, α, and 750' (at positions symmetrical to the center point of the center member 752, on the same side), and similarly form a bridge circuit.

In addition, when adding the γ force detection module 750 to the output rod 747, either the screw 751 alone is placed at the center of the output rod 747, or in this configuration, when torque is applied to the outer ring 753, the screw 751 may loosen. Therefore, it is actually necessary to engage the pin protruding from the center member 752 with the output shaft 757 to prevent it from rotating, and to fix it with the screw 751 at a position deviated from the center position.

Further, this also applies to the connection between the output shaft 757 and the support body 754.

With this configuration, the central member 752 is fixed, and when a torque around the central axis (γ axis) is applied to the outer ring 753, the leaf springs 750α, 750b, 750C, and 750d are deflected. This deflection is measured by strain gauge 754 (L, 754, 75
4t' and 754d, and the output is taken out via the bridge circuit, the torque T regarding the two axes (γ) is detected.
, r can be detected.

(b) Explanation of the configuration of the control section FIG. 4 shows the force control section 87, the combining section 85, and the servo control section 83.
and a detailed circuit diagram of the power amplifier 84. In the figure, the same parts as shown in FIG. 2 are indicated by the same symbols, and 800 to 803 are composite circuits for each axis,
Force control commands PF, r to PFr and CP from the force control unit 87
804 to 807 are dead zone parts that output the difference between the speed commands VJ to Vr from U82, and the dead zone @W3: to Wr from the CPU 82 is set, and the force measurement value F x =
It provides a dead zone for F r, and the force measurement value F
This is to perform feedback of a signal through a nonlinear element for x to Fr. The characteristics of the dead band sections 804 to 807 are that when the absolute value of the input is smaller than the dead band width setting value W, the output is zero, and on the other hand, when the absolute value of the input is larger than the dead band width setting value W, the input is a positive value. When , the output is (input - dead band width setting value), and when the input is a negative value, the output is (input + dead band width setting value). 808 is a switch, and is it a power meter? glJ value F x
809 is an analog/digital converter (referred to as an A/D converter); sensor 3
A device that converts the analog force meter side values FX, FY, FZ, and Fr from into digital values and outputs them to the switch 808.
83α to 83d are servo circuits, and the position detector s
5fasutDm reigning f1. ! Knee 7'm rank 11t
84α to 84d are power amplifiers, each of which controls the position based on the difference between CX, Cy, and CzCr, and the speed based on the difference between the commanded speed v'~V'7 and the actual speed. Drive current is supplied to the motors of each axis based on the outputs of circuits 834 to 8.1.

Therefore, when the switch unit 808 is turned off by the CPU 82, the output of the force sensor 3 (i.e., the A/D converter 80
9) is the dead zone 8o4゜805.806.807
There is no input to the CPU 8, the feedback is turned off, and the CPU 8
Command positions MCX, Cy, C according to the teaching data from 2
The Z, Cr, and U command speeds vx, vy, vz, and vr are input as they are to the servo circuits 83α to 83d- for each axis, and the positions and speeds are controlled.On the other hand, when the switch unit 808 is turned on by the CPU 82, is turned on, and the output of the force sensor 3 becomes the control output PFX, P via the dead band parts 804 to 807.
FY, PFZ, PFr and the synthesis section 800-80
3 and input the command speed vX.

, VY, VZ, Vr combined output V'X, V'
Y, V'Z.

The v'r force is given to the water circuits 83α to B3d as a speed command.

Note that there are two X-axes, an xl-axis and an x2-axis, or there are actually two X-axes, although this is shown by one X-axis servo circuit 83α and one amplifier 84α.

(C) Explanation of teaching data FIG. 5 is an explanatory diagram of teaching data in the memory 81. Generally, teaching data is described by motion (work) commands and position coordinates.

In this embodiment, the following macro instructions are used as operation instructions.

In the figure, reference numeral 81α denotes a work table in which work procedures of operations are described in sequence, and in this example, one micro-operation is described as a macro instruction that summarizes one work as described later.

Blb is a coordinate table that stores designated positions for macro instructions.

This macro instruction consists of a verb part and an object. In the example shown in the figure, Masuda is command rPIcKJ, and installation is command rPIcKJ.
PLAcEJ is the verb part, and "DISK" (meaning disk 9) and rSPINDIJJ (meaning spindle to be described later) are the objects.

For example, the sequence “100m rpIcKDISKJ
is a first movement operation to move to a specified point (described later), then a second movement movement to descend at a specified speed), a disk (DISK
This is a macro command to pick up )9 by suction (CPICK) (new work operation) and return to the specified point (third movement operation), and rPLAcg 5PINDLEJ of sequence "110".
The first movement operation is to move to a specified point), the second movement movement is to descend at a specified speed), and the spindle (
SPINDLE) (work operation) and return to the designated point (third movement operation).
In "IcK", the micro-movement commands are rMovE (move) J, r D OWN (descend),, IJ
.

rVAcUUM ON (adsorb) J, rMOVE (
It is necessary to consist of four commands such as "Move", which makes work instruction very troublesome. Therefore, such macro instructions facilitate work instructions.

The basic operations indicated by this macro command are: (1) Move to the specified point (first movement operation) ■ Descend from the specified point at the specified speed (second movement operation) Plane After stopping, work on the target member (work 0 In the above example, rPICK, 4 indicates retrieval as the operation in (2). rPLACEJ shows installation as work (2). Therefore, anything that can be performed using this basic operation can be applied to other tasks as well. For example, rs
cREWJ indicates screw tightening as [phase] work, rS
EARCHJ indicates searching for screw holes using a bottle as the [phase] work.

In addition to such assembly work, rFIRE is also available for welding work.
Welding is performed as the work (2) by the macro command J, measurement of the height of the target member is performed as the work (2) by the macro command rMEsUREJ for measurement work, and similarly, the macro rPAINTJ is used for painting work. rsEAL for painting or sealing work on command
Sealing is performed using the macro command J.

On the other hand, the coordinate table 81b stores the position coordinates of the aforementioned designated points of rDI SKJ and r 5PINDELEJ, which are the objects of the macro instruction words.

The advantage of this method is that when there is no change in the work procedure and the coordinate values change, it is only necessary to change the coordinates in the coordinate table 81b; conversely, when there is no change in the coordinate values and there is a change in the work procedure, l In this case, it is only necessary to change the work order in the work table 81α, and therefore, there is no need to recreate all the teaching data every time the work contents are changed, and such changes can be easily made.

Such teaching data is C; A D (Compt
[Lter kidtd])gziyrL) system, or by teaching the robot with a teaching box and then teaching the robot directly, and is stored in the memory 81.

In the second movement operation of the ID described above, as will be described later, the output (force measurement value) of the force sensor 5 described above is used to stop approaching, contacting, and pressing the target member, and positioning is performed. Therefore, one point is sufficient for the teaching position, which is 0 points.

(d) Explanation of operation of configuration of one embodiment FIG. 6 is an explanatory diagram of work operation of one embodiment of the present invention, FIG. 7 is a processing flow diagram for the operation, and FIG. 8.

FIG. 9 is an explanatory diagram of the operation.

In this embodiment, the rpI of the work table 81α in FIG.
CK DISKJ, rPLACE 5PINDLE
It shows the operation corresponding to rPICK DISK
J, the suction hand 5 moves from the current position A point to the B point specified by the coordinates of rDIsKJ in the coordinate table 81b, descends in the C direction, and suctions the disc 9 held on the jig 6 of the pallet 11α. hold, return to point B, then rPL
Suction hand 5 by ACE 5PINDLEJ
moves from point B to point D specified by the coordinates rSpINDLEJ of the coordinate table 81A, descends in the direction F, and spindle 1 attached to the base 7 on the pallet 11b.
FIG. 7 is a flowchart of the process for taking out the disc 9. FIG.

(2) When the 5TART command is input from the operation panel 80, an operation start command is given to the CPU 82 via the bus 89.

As a result, the CPU 82 decodes the command statement in the work table 81α of the memory 81 and reads rpICK DISKJ.
If so, it is determined that it is suction removal (PICK), and the sedimentation process is started.

Next, the CpU 82 reads out the rpIsKJ column of the coordinate table 81α in the memory 81, and reads the above-mentioned point B (x2, y,
z), and position commands CX2, 'CY
, CZ to the servo control section 83 via the bus 89, and x, y, z speed commands VX2, VY, to the synthesizing section 85.
Give VZ.

As a result, the power amplifier 84 is controlled by the servo control section 83.
Drive current SX2. SY and SZ are X-axis motors 10α
, Y-axis motor 24 and two-axis motor 26.

Therefore, the X-axis motor 1 of the X-axis (X2-axis) module 1α
0α, Y-axis motor 24.2-axis motor 26 is driven,
The vacuum suction hand 5 picks up the pallet 1 on the X-axis module 1α.
1α, the jig 6 is positioned at the designated point B (FIG. 6) on the disc 9.

■Next, the CPU 82 checks the current position of each axis of the hand position detection circuit 88 [PX 2 , PY , PZ , etc.] to determine whether or not the hand has reached a predetermined position, and after stopping at the predetermined position. After the specified time has passed (about 0.5 seconds)
, it is assumed that the vibration of the force sensor 3 has stopped, and the force feedback is turned on, that is, the switch 808 of the force control section 87 is turned on, and the output PFX-PFZ of the force control section 87 is sent to the synthesis section 8.
5 can be input. Further, a valve of a vacuum pump (not shown) is opened to enable suction by the suction hand 5. Note that the vibration stop may be determined using the signal from the force sensor 3 based on the magnitude of the vibration amplitude.

(2) The CPU 82 provides the speed command value VZ to the synthesis section 85 via the bus 89. As will be described later, since the force command PFZ is O before contact with the disk 9, it is outputted as a command speed to the servo circuit 83C, and the speed of the two-axis motor 26 is controlled.

Therefore, the suction hand 5 descends (approaches) toward the disk 9 at the commanded speed vl.

On the other hand, the CPU 82 monitors the two-axis current position FZ of the hand position detection circuit 88 via the bus 89, and when FZ shows the same value for a certain period of time, determines that the suction hand 5 (that is, the two axes) has stopped.

During this time, external force adaptive control based on the output of the force sensor 3 is executed autonomously. This will be explained with reference to FIGS. 8 and 9.

When the hand 5 is approaching the disk 9, the force measurement value F of the force sensor 3
Since z is zero, the output V' z of the two-axis synthesis circuit 802
"Vz, and the two-axis motor 26 is connected to the servo circuit 83.
The speed is controlled by C via the power amplifier 84C, and the hand 5 is lowered.

On the other hand, the hand 5 is placed on the disc 9 at a position Pt as shown in FIG.
When contact is made at (time tt), the force sensor 3 is deflected because it is composed of a leaf spring, and the strain gauge of the force sensor 3 that detects this deflection generates a force measurement value Fz. The attachment continues to move at the speed given by the command Vz until the force (force measurement value Fz) is reached, and at time t2 it moves to Fz.
By satisfying z:>Wz, the force command P F z
occurs.

Therefore, the output V' z of the two-axis synthesis circuit 802 is (V z
-PFz), and the composite speed command V'z is the speed command V
The force command PFz is subtracted from z, and the speed command value appears to be smaller, and the node 5 begins to decelerate.

, the parallel plate spring of the force sensor 3 is further deflected by the lowering of the hand 5 due to the rotation of the two-axis motor 26, the force measurement value Fz increases, and the hand 5 is further decelerated.

In the final equilibrium state, the two-axis motor 26 stops at the position PO at time to, and the actual speed value becomes zero.

At this time, input command Vz and force command PFz are equal, and force command P
Since Fz is pressed against the object 9 by the deformation of the force sensor 3, it becomes a force FO, so the input command Vz moves as a force command.

Therefore, the magnitude of the input command Vz controls the pressing force.

The pressing force at this time is the sum of the force command value based on the input command Vz and the force corresponding to the dead zone + [W.On the other hand, the speed during approach depends only on the input command Vz, so the moving speed and For pushing, the force can be controlled independently 0 Since the human power command Vz serves as both the speed command value during approach and the force command value for pushing when force is generated, the output of the speed measurement value Vz and force measurement value Fz Depending on the relative size of the levels,
There is a possibility that one input Vz cannot satisfy both the optimum moving speed and the optimum pressing force at the same time.

For this reason, an input is provided to make the feedback amount of the force measurement value Fz variable, that is, the speed command Vz and the dead band width Wz are set independently according to the dead band width Wz, and the optimal speed command value and force command value are set. and can be obtained with one input command. In addition, since the feedback gain of the force measurement value Fz can be fixed, the pressing force can be applied over a wide range while maintaining the loop gain as a closed loop control system at a constant l· (that is, while guaranteeing the stability of the control system). Things that can be changed.

And when the hand 5 is floating in the air (that is, when the hand 5 is not pushing another object) and you want to set the input to zero and keep it in place, there is an offset fluctuation in the force measurement value of the force sensor 3. Also, is it possible that the position of the JS hand 5 drifts? In this embodiment, it becomes insensitive to offset fluctuations smaller than the dead band width, and the drift phenomenon disappears.

Therefore, while the shock at the time of contact is absorbed by the deflection (deformation) of the force sensor 3, the rotational speed of the motor 26 is continuously reduced, and when the speed reaches zero, appropriate pressing generates force. In this way, the three processes of approaching, contacting, and pressing are performed continuously, smoothly, and autonomously.

In other words, when a robot works to grind an object, the robot's hand approaches the object, makes contact with it, and then presses the object with a certain force to grip it. , the hand gripping the object approaches the object, makes contact with the object, and then presses the gripped object against the object with a constant force to fit and attach the object.

In the present invention, the above-mentioned three processes of approach, contact, and constant force generation of the hand, which is the working part, can be controlled continuously and autonomously.

In other words, since the rotational speed of the motor decreases continuously before and after contact, the impact at the time of contact can be kept small, and the linear force output value of the force sensor 3 is used to change the timing from speed control mode to force control mode. You can switch easily and smoothly. In addition, when pressing is under force control, the feedback of the speed measurement value (actual speed) is considered as a so-called damping ring of state variable feedback, so the control system is extremely stable. When the depth at which the hand 5 picks up the magnetic disk 9 or the spacer 90 changes, such as when picking up a stack of magnetic disks 9 and spacers 90 one after another and moving it to another location. However, since the desired hand descending speed and desired force can be obtained automatically and always, teaching the distance in the depth direction is no longer necessary, and the burden on humans to teach the robot the work is reduced. This will make the robot lighter and take one step forward in making the robot more intelligent. Also, at this time, as shown in Figure 9 (8), there is a relative positional deviation in the Y-axis direction with respect to the Y coordinate of the teaching data of point B, and the hand 5 If the force sensor 3 contacts the jig 6 while approaching the disc 9 and the fitting is not performed smoothly, the contact with the jig 6 causes the force sensor 3 to
The Y-axis force measurement value Fy is generated as shown in Fig. 8 (Q) and F
! >>WJ, the force command PFy is outputted, and the composite speed command V'y (=-P
F31) is generated, the Ym motor 24 is driven via the Y-axis servo circuit 83b and the 1° warp amplifier 84b, and
The relative position error in the Y-axis direction is corrected so that the hand 5 automatically moves in the direction of the arrow Y in the figure, which is the external force applying direction, and the hand 7 fits into the jig 6.

Similarly, as shown in FIG. 8(B), there is a relative positional deviation in the X-axis direction with respect to the teaching data X2 coordinate of point B, and the hand 5 comes into contact with the jig 6 while approaching the disk 9. Even if the fitting is not carried out smoothly, the Y-axis force measurement value Fx is generated from the force sensor 3 due to contact with the jig 6 as shown in FIG. 8 (Q).
Similarly, the X-axis motor 10α is driven via the X-axis servo circuit 83α and the power amplifier 84α, and the pallet 11α, that is, the jig 6, moves in the direction of the arrow in the figure, causing a relative position error between the hand 5 and the jig 6 in the X-axis direction. will be corrected.

After all, in the range where the maximum drag is smaller than the dead band width, the relative position error between the hand 5 and the jig 6 is absorbed by the deflection of the force sensor 3 while following the jig 6. On the other hand, when the maximum drag exceeds the dead band width, the motor The relative position error is corrected by driving, and the center of the jig 6 (the center of the disk 9) and the center of the hand 5 are automatically aligned, and fitting into the jig 6 is executed.

■ When the hand 5 is fitted into the jig 6 in this way and the Z-axis is further stopped, the CPU 82 monitors the output of a pressure sensor (not shown) that detects the negative pressure in the intake tube via the bus 89, and It is detected whether the disc 9 has been attracted.

When the suction hand 5 is fitted with a slight inclination to the jig 6,
A gap is created between the suction surface 50 and the disc 9, and the disc 9 cannot be suctioned by the above-mentioned predetermined pressing even if the force Fz is applied. Therefore, in order to relieve such a situation, the following relief operation is performed when the adsorption is not performed.

First, the CPU 82 makes the dead band widths of the Y-axis and Y-axis dead band portions 804 and 805 zero through the bus 89.

That is, since the inclination is due to the posture of the hand 5 in the X and Y directions, the maximum drag in these directions is set to zero.

Next, the CPU 82 increases the speed command Vz by α in order to increase the pressing force on the Z axis, and accordingly drives the Z axis motor 26α. Then, as described above, it is determined whether or not the adsorption is performed. If the adsorption is not performed, the speed command is further increased by α, and the 2-axis motor 2
6α is driven. Repeat this multiple times (10 times in the figure)
, If it still does not attract, give up on the attraction operation, set error number 1, and proceed to step ■. On the other hand, if it is determined that it has been adsorbed, immediately proceed to step ■.

■ Assuming that the adsorption occurred in the above-mentioned step ■ regardless of the rescue operation, the CPU 82 then detects the X of the hand position detection circuit 88,
Read the Y coordinates Px, Py via the bus 89;
rl) ISK of the coordinate table 816 in the memory 81 via
Rewrite the X coordinate (here, the X2 coordinate) and Y coordinate in column J. As a result, the result of absorbing relative positional deviation due to external force adaptive control can be reflected in the teaching data, and the coordinates of point B are replaced with coordinates with no relative positional deviation.
Next, when steps ■ and ■ are completed, the CPU 82
The shaft speed command value VZ is set to zero, the pressing force is released, and then the aforementioned switch 808 is turned off to turn off the feedback, or the dead zone width is made sufficiently large.

Furthermore, the CPU 82 retrieves the coordinates in the rDIsKJ column of the coordinate table 81h mentioned above in order to raise the suction hand 5, and
Return to point B in the same manner as step ■.

As a result, in order to fit the removed disk 9 onto the spindle 12, the next sequence rpLAcj(SPl, NDLEJ) of the work table 81α is performed as described above. rsPIND of the coordinate table 81b.
Continuously record the coordinates in the LEJ column, descend in two directions in the same manner as steps ■, ■, and ■, approach the spindle 12,
After contacting and pressing and fitting with the spindle 12,
The suction is released and the disk 9 is attached to the spindle 12. At this time, the relative position error is corrected as in step (2), and the process returns to point D in the figure as in step (2).

In step (3), the teaching data indicated by the absolute position can be corrected by the operation result obtained by performing state adaptive control based on the teaching data and correcting the relative position shift. In this way, the macro instruction can be Even if the work command used is used, the teaching position only needs to be one point, and the hand 5 can move the force sensor 3 by simply instructing the speed without teaching the position as in step (2).
Utilizing the output of , it automatically stops when it comes into contact with the disc 9, and the positioning 0 becomes possible.Therefore, the suction operation of the disc 9 becomes possible.

Similarly, the mounting operation is automatically stopped and positioned by fitting and contacting the spindle 12.

Further, even if there are a plurality of discs 9, the work of taking out and attaching the number plate can be carried out in the same manner without teaching the position.

Furthermore, even if the X and Y coordinates of point B are not accurate, the relative positional deviation is corrected or absorbed by the output of the force sensor 3, making teaching even easier. Furthermore, by correcting the teaching data as soon as step (2), the next operation can be performed more smoothly.

(-) Description of Other Embodiments FIG. 10 is an explanatory diagram of processing of the CPU 82 as another embodiment of the present invention.

In this example, the functions of the synthesis section 85 and force control section 87 are
This is done by executing the program of U82.

In the main routine, the CPU 82 analyzes commands in the teaching data, executes them to create position commands, speed commands, and force commands for each axis, and sends them to the servo control unit 8 via the bus 89.
3, and outputs the current position Pr to P from the position detector 88.
The position of each axis and the force measurement value F r =F x from the force sensor 3 are monitored by x2. In the normal feedback off mode, the command position and command speed are directly given to the servo control unit 83 as shown by the dotted line of the switch, and each axis is moved to the command position via the power amplifier 84. On the other hand, in cafe feedback on mode,
An interrupt processing routine for feedback control is executed at a predetermined period. That is, the force meter side value of the force sensor F r = F
x is offset corrected, further processed as a dead zone, and multiplied by the device gain to obtain the control output, which is subtracted from the commanded speed or commanded force as the commanded speed V'r-V'3:2, as shown in the switch solid line. Function generator 83 via </size 89
In this dead zone process, the dead zone width W r −W x
can be set freely, and when the set dead band width is W, if the force measurement value F is positive, (FW) is output, and if it is negative, (
F+W).

This feedback control enables the following control as described above.

When a commanded speed Vz is given to a certain axis (for example, the Z axis), the object moves at the commanded speed Vz while approaching the article. .

On the other hand, when the object is contacted at point P1, the force sensor is bent and deformed, and a force measurement value Fz corresponding to the amount of deformation is generated, but within the range of the dead band width W, the control output Pz is zero, so the speed remains Vz. , Fz>W, a control output Pz is generated and the command speed Vz is decreased. Therefore, the moving speed of the two axes also decreases, and when P z = W z, the z axis stops. At this time, force is being applied to the article by pressing Pz, and therefore the command speed Vz becomes the command force. There is. This allows the article to be approached, touched, and pressed continuously.

This can be used to adsorb and insert the aforementioned items, adsorb screws,
Axes that allow smooth tightening of screws, etc., and for which no position command or speed command is given (for example, X, Yd)
When an external force is applied in the axial direction of the shaft, a control output is generated based on the force measurement value F due to the deflection of the force sensor, which becomes a command speed, and the shaft is driven in the direction in which the force measurement value F becomes zero. By utilizing this fact, even if the hand and the object come into contact and the fitting becomes difficult when fitting the hand and the object, the force sensor output is automatically driven in the direction of zero, ensuring a smooth fitting. This makes it possible to adsorb and insert objects even more smoothly.

■ Description of another embodiment In the above embodiment, the force feedback by the force sensor 3 was explained as an example, but the state feedback may be by other contact sensors (for example, a microswitch), and the proximity sensor, distance sensor, etc. Status feedback may be provided by a non-contact sensor such as a sensor. Furthermore, the content of the work is not limited to taking out objects, but may also include welding to objects, etc. In this case, the working part will be a welding torch, and the work part will be stopped without contact before contacting the target member 9. It is better to change the gain α to

Further, the hand 5 is not limited to a vacuum suction hand, but may also be an electromagnetic suction hand, a gripping hand, a hook-shaped grip hand, an electric screwdriver, etc., depending on the content of the work. You can choose as appropriate.

Further, the macro command may also be the basic operation (I) (ID) as one macro command, and (IID (IV)) as the second macro command.

If the aforementioned servo control section 83 is made up of a function generation section and a closed loop control section proposed by the present inventors, a more stable servo system can be realized. This servo circuit is described in the magazine "Nikkei Electronics 1981", for example.
9.28'' (published by Nikkei McGraw-Hill), etc., so it will not be described in detail.

Furthermore, although the robot has been described as an orthogonal robot with a divided X axis, it can also be applied to other orthogonal robots, and may even be a scalar articulated robot. Although the explanation has been made using assembly as an example, the present invention is not limited to this, and other operations may be used.Furthermore, the movement command in the Z-axis direction is not limited to a speed command, and may be a position command.

In this case, it is necessary to specify the coordinate value in the z-axis direction below the pallet 11b, but for example, after teaching the teaching point B, it is necessary to specify only the coordinate value in the z-axis direction using the operation panel 80 or the like. All that is required is to create updated teaching data. Therefore, the number of teaching points can be substantially reduced to one.

Although the present invention has been explained above using Examples, the present invention can be modified in various ways in accordance with the spirit of the present invention, and these are not excluded from the present invention.0 [Effects of the Invention] As explained above, According to the present invention, even if a macro instruction indicating multiple operations is introduced as a work instruction, only one teaching position is required, making the teaching extremely easy, and further demonstrating the effect of introducing the macro instruction. .

Therefore, it becomes easier to teach complicated tasks, and it becomes possible to have automatic working devices easily execute complex and highly accurate tasks, which greatly contributes to the promotion of work automation.

[Brief explanation of drawings]

Fig. 1 is an explanatory diagram of the principle of the present invention, Fig. 2 is an overall configuration diagram of an embodiment of the present invention, Fig. 3 is a block diagram of the force sensor configured in Fig. 2, and Fig. 4 is an essential diagram of the configuration in Fig. 2. FIG. 5 is an internal configuration diagram of the memory configured in FIG. 2, FIG. Figure 7 is an explanatory diagram of the operation in the processing flow, Figure 9 is the
FIG. 10 is an explanatory diagram of another embodiment of the present invention, and FIG. 11 is an explanatory diagram of automatic work operation. In the figure, 3... Status sensor, 5... Working part, 8... Control device, 9... Target member, MT... Moving means, CT... Control unit,

Claims (3)

[Claims] (1) In an automatic working device in which a moving means moves a working part that performs work on a target member within a work space and causes the working part to perform work on the target member, the state of the working part with respect to the target member is checked. a state detection means for detecting; a memory that stores a work command written in a macro command indicating a work including at least a plurality of moving actions; and a specified coordinate position corresponding to the written work command; and a work command in the memory. a first movement operation that controls the movement of the movement means to a coordinate position corresponding to the work command; , after the first moving operation, a second moving operation that controls the movement of the moving means based on the moving command and the detection output of the state detection means to move the working part in the direction of the target member to perform the work; An operation control method for an automatic work device, characterized in that: (2) After performing a second movement operation in which the control unit moves the working unit in the direction of the target member to perform the work, the moving unit is controlled to move to the coordinate position of the coordinate table, and the working unit An operation control method for an automatic working device according to claim (1), characterized in that a third moving operation is executed to return the . (3) The macro instruction indicates a working operation of the working section, and the control causes the working section to perform the working operation according to the working instruction after the second moving operation. An operation control method for an automatic working device according to paragraph (1) or paragraph (2).