JPS5979312A - Controlling device of direct teaching type industrial robot - Google Patents

Abstract

PURPOSE:To eliminate unstableness of control by feeding back difference information of positional information from a detector provided in the driving source side and positional information from a detector provided in the robot element side through a transmission means, to input side of a controlling system. CONSTITUTION:Positional information thetan of a driving source motor M is detected by a detector 81 and positional information thetaL of robot element that passed through a reducer 41 is detected by a detector 49. Data of the detector 49 is calculated by a counter 203, and calculated value thetaL is fed back FB to an arithmetic device 204 and position control device 101. At the same time, difference of thetam-thetaL of detected data of detectors 81, 49 is calculated by a difference counter 202, and data obtained by multiplying this by a constant KF is fed back to the device 101. The device 204 stores thetaL in a memory 206, and at the same time, sends target position data thetain to the device 101. The device 101 composes a value obtained by multiplying the difference of thetain and thetaL by proper gain K1 and Kp (thetan-thetaL) and outputs a speed command signal omegain. The signal omegain is D/A converted and amplified, and the motor M is controlled.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a control device for eliminating control instability in a direct teaching type industrial robot, and in particular to a control device for eliminating control instability caused by the intervention of a transmission means such as a speed reducer. The aim is to ensure the same stability as a remote control teaching robot by feeding back the amount to the input law of the control system.

In general, teaching methods for industrial robots can be broadly divided into two methods: remote control teaching methods, in which the operator operates switches on a remote control box to move the robot in an inching operation; There is a direct teaching method (referred to as manual teaching) in which teaching is performed manually by direct operation.The latter method has the advantage of shortening the teaching time compared to the former method, but it is also designed to allow easy manual teaching. Therefore, a mechanism for lightening movement by inserting a clutch mechanism or the like between the drive source and robot elements such as arms is required.As a result of providing such a lightening mechanism, the position detector of the robot element becomes It must be installed after the clutch, that is, on the robot element side (if a position detector is installed on the drive source side such as a motor, as in a remote control system, the clutch will be turned off.
In other words, position data cannot be captured at the time of teaching. ). The normal robot elements are the drive source and the robot 5.
A transmission means such as a speed reducer is installed between the robot element and the robot element, and since such transmission means has control delays due to gear backlash, twisting, deflection, etc., a position detector attached to the robot element side is used. If the output from the loop is used as the main feedback amount, vibrations from the transmission means will enter the control loop, making the control unstable, so if you increase the loop gain to speed up the control, it will cause oscillation-like deafness. It had some drawbacks.

The purpose of the present invention is to eliminate the instability in the control of direct teaching robots as described above. By adding
The aim is to achieve stable control similar to the remote control teaching method that uses feedback from the drive source side, while employing feedback from a position detector attached to the robot element side.

Next, embodiments embodying the present invention will be described in detail with reference to the accompanying drawings and the first page.

FIG. 1 shows an example of a direct teaching type welding robot to which the present invention can be applied.

The swivel table 301 has a drive unit such as a motor speed reducer and a clutch therein, and rotates in the direction of θ1. The arm and wrist drive sections 302 are surrounded by a cover, one on each side of the robot, and one arm drive section and one wrist drive section built into each. That is, the first arm motor 314 attached to the side surface of the moving part 302 rotates the first arm link 313a through a reduction gear and a clutch, as will be explained later in detail with reference to FIG. Directly above the river motor 314, there is a shaft that can freely swing around the support 10! The arm 304a that is integrally attached to the 1S1 arm 304 and the first arm link 313
a and a link 313b connecting these arms 3133 and 304a form a parallel link, and the link 31
As the arm 304a rotates, the arm 304a and the first arm 3
04 rotates by the same angle σ2 in the same direction as the arm 313a. The second arm 305 is swingably attached to the 1ilhll attached to the tip of the first arm 304,
The rear end 12 of the second arm 305, the link 14 which is rotatably driven by the second arm motor installed on the back side of the drive unit 302 (the back side of the motor 314 in the figure), and the shaft 1°. Link 13 rotatably supported as a center
and a weight link 3 connecting the links 13 and 14.
16, a connecting link 303 that connects the link 13 and the rear end portion 12 of the second arm, and the first arm 304 constitute a second parallel link. The second arm 305 is rotated in the direction 03 by the rotation of the second arm motor through two sets of parallel links.

Furthermore, regarding the wrist, a swinging motor 315 is attached to both left and right sides of the base of the first arm 304, and a swinging motor provided on the back side thereof, and these wrist driving motors move the first arm and the first arm. The chain built in the second arm rotates 111iL, and the end of the chain reaches the wrist mechanism section 306, giving the wrist 306a a swinging motion in the 04 direction and a turning motion in the θS direction, and the end of the chain reaches the wrist 306a. Welding torch 308 attached to
The robot guides the user to the desired position required for welding operation.

Next, a gravity balance mechanism for offsetting the gravitational moments of the first arm and the second arm will be explained. First, regarding the first arm, there is a link 13 rotatably supported at the lower end of the first arm 304 around a shaft 13, a link 14 parallel to this, and a vertical wave link connecting the links 13 and 14. 316 forms a parallel link,
The upper end B of the wade link 316 is on the vertical line of the connection point between the link 13 and the knee link 316, and between the tip B and the rear end A of the second arm, there is a tension spring or the like. A spring balance mechanism 310 consisting of the following is tensioned, and this gravitational moment in the form of the triangle ABD is canceled out. Note that the tensile force of the spring balance mechanism 310 can be balanced regardless of the value of the inclination angle θ2 of the first arm if the attachment is configured so that it is proportional to the distance between AB, which is the length. Regarding the second arm, the gravitational moment around the axis 11 of the second arm 305 is offset by the weight of the spring balance mechanism 310 and the weight of the wave link 316 and the like. By using such a gravity balance, no force is required for the operation during teaching, and as will be described later, the accuracy of position control is improved.

As an example of the structure, a wrist drive mechanism will be described with reference to FIG. 2. This type of drive mechanism is not only used for wrist movement (it can also be used for arm movement.'! In Figure 52, M is a DC servo motor for driving, and the motor M is connected to the side opposite to the output shaft. A pulse encoder 81 that generates pulses according to changes in the rotation angle of the motor M, and a tacho generator 80 as a speed detector are attached to the output shaft side.A concentric type motor is installed on the output shaft side to obtain a large reduction ratio. A clutch 51 is attached to the output shaft of the reducer. This clutch is a shield tooth clutch, and the clutch 56 is the claw of the clutch. By engaging or disengaging the parts 57, the rotation of the output shaft 42 of the reducer 41 is transmitted to the robot a element drive shaft 59 (ON) or shut off (OFF).
be done. The structure of the clutch 51 is such that a rotor 52 surrounding a coil 53 is fixed to an output shaft 42, and a rotor 52 surrounding a coil 53 is fixed to an output shaft 42.
The coil 53 is fixed to the case 54 of the drive mechanism without contacting the rotor 52. Robot element installation is on axis 5
When the coil 53 is excited, the armature 55, which is spline-fitted to the armature 9 and slidable in the axial direction, is drawn toward the rotor, and the pawls 56 and 57 engage with each other. Conversely, when the coil 53 is not energized, it is moved in the opposite direction to the coil 53 by a spring (not shown), and the claw 5
6.57 engagement is released. Furthermore, in order to attach the robot element, a large gear 60 is coaxially engraved on the shaft 59, and the large gear 60 meshes with a small 9M wheel 63 coaxially integrated with a shaft 61 rotatably attached to the casing. Furthermore, a large gear 64 coaxially integrated with the shaft 61 is a pulse encoder 49.
It meshes with a small gear 66 attached to a drive shaft 68 . The robot element drive shaft 59 itself has the structure of a sprocket, and is connected to the wrist portion by means of a jaw as described above. In the case of an arm, a portion of the parallel link described above is directly attached to this drive shaft 59. The pulse encoder 81 is connected to the drive source 1 (!!
The pulse encoder 49 is a component of the first position detection means for detecting the position of the robot element.

Therefore, when the motor M rotates, the Teso output 1Fil 142 rotates at a low speed through the deceleration 'm, 41, and the clutch 51
When is in the ON state, the rotation of the output shaft 42 is directly transmitted to the robot four-element drive shaft 59. And robot Shosu! ! 1
The rotation of the moving shaft 59 is transmitted to the pulse encoder 49 via gears 60, 63, 64°66.

Gear trains 60, 63, 64, and 66 are inserted in order to equalize the number of output pulses of the pulse encoder 41 and the pulse encoder 49 on the motor ('Q-<Bu]i%) side.

That is, when the reduction ratio of the reduction aIS 41 is, for example, 300, the speed increase ratio of the gear train is 30, and the number of pulses generated per one rotation of the shaft 68 of the pulse encoder 49 is 100 pulses, the pulse encoder 49 rotates one rotation. 1 hit
It is sufficient to use one that outputs 000 pulses. By aligning the numbers of output pulses from both pulse encoders in this way, subsequent calculations become easier. However, a branching device or the like may be used to equalize the number of pulses. In this way, by integrating the respective pulse numbers of the position detectors 81 and 49 attached at two locations, it is possible to obtain the count values HI and θL corresponding to the rotation angle of the motor shaft and the rotation angle of the arm. In a steady state, θm and θL have a one-to-one correspondence.

Next, FIG. 6 shows a block diagram of the entire control device according to the first embodiment of the present invention. As shown in the figure, this control device generally uses target position data θi outputted from an arithmetic unit 204 constituted by a CPU of a microcomputer.
Difference between the value of n and the value of the counter 203 indicating the position IN of the arm which is a robot element, that is, θL (θin - θL)
is multiplied by an appropriate gain by 1 to output the speed command ωin, convert the speed command ωin into an analog quantity with a DAJ converter, rotate the motor M via amplifiers 201 and 103,
A normal proportional control method is adopted in which the robot elements are rotated by the rotation of the motor via the reduction gear 41. The rotational speed ωrn of the motor is detected by a tacho generator 80, multiplied by a constant value by an amplifier 205, and fed back to the input side of a motor driver circuit 103, which is an amplifier. The position detector 49 employs an incremental type pulse encoder. In this case, it is necessary to install a counter 203 as an integrator in order to calculate the absolute value of the rotation angle of the arm. When such an incremental pulse encoder is used, the position detector 49 corresponds to the second position detector.

In the present invention, in addition to the conventional control means as described above, the first position detector 81 as described above is provided on the motor side, and furthermore, the first position detector 81 as described above is provided on the motor side. A difference counter 202, which is a type of conversion means, is provided for integrating pulse signals and simultaneously counting the difference value. In this case, the first position detector 81 also uses an incremental pulse encoder. However, it is also possible to use absolute type pulse encoders as the first and second position detectors, in which case the counter 203 is omitted and the difference counter 202 simply calculates the difference between the output values from both position detectors. It is fine as long as it is done. In the case of the embodiment shown in FIG. 6, the difference counter 202 feeds back the position difference information between - and θL (鵠−θL) to the position control device @101 in the same way as the position information θL of the robot element. This position control device 101 may be configured as a part of the calculation device 204. Within the position control device 101, these two signals are added together and compared with the taught position data θin from the arithmetic unit 204 as position information θi. If you simply add the same rank (δ difference information - θL and position information record, the sum θt is %1
is equal to For this reason, even though θL is set as a position (17 reports), - is apparently controlled as the main feedback amount, and vibrations and torsions of the mechanical system caused by transmission means such as reducers are detected from within the control loop. etc. can be excluded.Also, an appropriate value KF (KF>1) for the difference information destination, -〇L
The control performance can be further improved by multiplying by 0L and adding it to 0L, but this point will be explained in detail later.

Furthermore, the relationship between θIII and θL does not match when the clutch 51 is OFF, that is, when the teaching operation is being performed, because they are unrelated, and even when the clutch 51 is reconnected during playback, the values are different from the beginning. becomes. In this case, it is not possible to accurately measure 0rn-θL during regeneration, so the value of the difference counter 202 is reset when the clutch 51 is turned on, that is, during regeneration. Before this difference counter is reset, the difference between 1 and θL becomes 0, and the subsequent value of -0L represents the amount of torsion, vibration, etc. of the reduction gear 41 and the like. The timing of generating this reset signal is preferably 11) at the time of small start, more precisely just before starting the reproducing operation. This is because the speed reducer 41 is twisted and θ□ and θL no longer match when the vehicle is accelerating. However, such a reset operation is not limited to the reset operation that completely sets the value of the difference counter to 0 as described above, but also the value of the difference counter 202 that occurs at the end of teaching or immediately before the start of playback is stored as the initial difference. A similar reset effect can also be obtained by subtracting the above-mentioned initial difference from the device difference information constituted by a difference counter at the time of reproduction and inputting it to the position control device as true position difference information. In addition, when absolute type encoders are used as the first and second position detectors, W(2
A reset means is used, such as transferring the value of the position detector to the first position detector.

Also, when teaching, the clutch 51 is turned off and the robot element is moved manually, and the position at that time is detected by the second position detector 49 and the counter 23, and the values are sequentially stored in the storage device 206 as the above-mentioned teaching position data and reproduced. At the same time, the teaching position data θin is taken out sequentially and stored in the counter 20.
The difference between the value obtained by combining the position signal θ1− from 3 and the position difference information from the difference counter 202 and the taught position data θin is calculated, and the value is multiplied by 1 by an appropriate gain and driven as the speed command signal ωin. output to the system.

A block diagram of such a control device as shown in FIG. 7 is shown in FIG. If the KF of the control device shown in Figure 6 is set to 1, it becomes the same value as the block diagram shown in Figure 8, which means that the control characteristics of a direct teaching robot have improved to the control characteristics of a remote control robot. What was done was understood.

Now, the case where KF becomes larger than 1 will be explained by modifying the expression of the transfer function using Laplace transform. First, the transfer function of 1, that is, the mechanical system, written in FIG. 7 etc. becomes the transfer function of the second-order lag. Here, JL is the inertia of a robot element such as an arm, K is a spring constant of a reducer, etc., and D is a viscous resistance that is the sum of the viscosity of oil or grease in the reducer, air resistance of the arm, etc. here 5
The coefficient of the term 2 is 11, /I (whereas the coefficient D/K of the term S is
However, for ordinary robots, the value is quite small. This shows that the transfer function (() is in a state where it is easy to oscillate. Also, the body, which is the transfer function of the motor, is generally expressed as Gm = T? - Tm5, but it is simplified here. Therefore, the lungs are assumed to be included in 1, and since the mechanical time constant of a servo motor with good characteristics and a tacho generator applying speed feedback is extremely small, it can be approximated by Assuming it to be zero, Gm = '...(2) will be used.

In the case without position feedback in Fig. 7, that is, the first
When Figure 0 is transformed, it becomes as shown in Figure 11.

If the transfer function of the body in Figure 11 is G, then substituting equations (1) and (2) into this and rearranging...(3) is obtained.

From equation (3), this means that ■5 has become. This is called q2 to mean an improved 1. Gi is also a second-order lag, but its natural frequency ω2 and viscous drag coefficient τ' are...(5)
It is expressed as

If D is considered to be approximately 0, τ' can be increased by 1·KF while τ is approximately O. It is said that about τ-0.7 is good for general servo mechanisms, and Kl・K
By choosing P appropriately, it can be -07.

If the difference is sufficiently small, the unique linear motion number ω1 does not change much.

Looking at the above results in terms of step response characteristics, an example in the case of ~.>1 is shown in FIG.

FIG. 9 shows the step response when KF=l, and it can be seen that the control characteristics are further improved when compared with this. This is because τ, the viscous drag coefficient, and the numerical value related to damping have increased due to the mechanical engineering, producing the same effect as when a damper for vibration reduction is attached to a robot arm or the like.

Also, in the transformation of the transfer function equation, deformation of the mechanical system, ie, the reduction gear, etc., due to the gravitational moment of °r-m was not taken into consideration, but in reality it does not have much influence.

For example, in the formula, deformation (twisting) of the mechanical system occurs during acceleration, and when this deformation occurs, the output to the motor is suppressed, but when a gravitational moment is actually acting on the robot, the arm There is a risk that the mechanical system will deform even though the arm is stopped, and control will be performed as if it were accelerating.In this regard, as shown in the example above, in the slave device, the mechanical system is deformed when the arm is stopped. Since the gravitational balance device i,'(-C is used to offset the gravitational moment that causes deformation in the system), the above-mentioned interval λ■ does not occur.

In other words, in a state where there is no gravitational balance, if the equation is 1, the error due to the deformation of the mechanical system will be magnified and the system will enter the control system, causing Lee to be accelerating even though it is at rest. Since such control is performed, it will have a negative effect on positioning control, but if the gravity balance is maintained, theoretically such errors will not occur, and the accuracy at the time of stopping will also be improved.

FIG. 13 shows a second embodiment in which the position detection circuit of the embodiment shown in FIG. 6 is modified. In this case, pulse counters 104 and 105 are installed in the position detectors 81 and 49, respectively, and a subtracter 106 that outputs a position difference (Ti signal) between both pulse counters 104 and 105 is used as a conversion means. In addition, after converting the position difference information into an analog quantity by the l)/A converter 107, the input terminal 911 of the motor drive circuit 103 passes through the amplifier 108 that provides a gain of Kp.
] is giving feedback. Therefore, in this case, the result is equivalent to the first embodiment shown in FIG. However, the pillars in this figure may differ from those in the embodiment shown in FIG. 6 depending on the gain in the position control circuit. It goes without saying that the present invention is also applicable to robots using hydraulic pressure or other means as a driving source.

As described above, the present invention provides a control device for a direct teaching type industrial robot in which a drive source drives a robot element such as an arm through a transmission means such as a speed reducer. a second position detecting means for detecting the position of the robot element and feeding back the position information to the input side of the control system;
From the position information obtained by the 20th place 11≦1 detection means, the difference in position between the drive source and the robot element is detected (iLWt, difference information is calculated and fed back to the input side of the control system. and converting means to convert the second digit to 1 during teaching.
iI. A position storage means for sequentially storing the teaching position data of the robot element sent from the detection means, and a reset for re-clearing the difference in position between the first and second position detection means that occurred at the time of teaching. and a control means for multiplying a signal of the difference between the taught position data and the composite value of the taught position data retrieved from the device storage means and the position difference information at the time of reproduction by a constant gain and outputting the result to the drive system. Since it is a direct teaching type industrial robot, it has a large damping effect, and the vibration of the 3A mechanical system due to the transmission means is suppressed, greatly improving the control characteristics of the robot for direct teaching. In addition, when the control method of feedback of the difference between two position detectors according to the present invention is combined with the gravity balance device, the problem that occurs when the gravity balance device is This solves the problem that despite the improvement in control characteristics, the positioning of the tip of the robot element decreases by 7J degrees, making it possible to improve control characteristics without changing position accuracy. Therefore, when feedbanking only the position information from the position detector installed on the robot element side, as in the case of a conventional direct teaching type industrial robot, the control system is shown in a block diagram as shown in Figure 3. If the loop gain Kl is small, a step response with minute vibrations as shown in Fig. 4 will occur, and if K is increased, an oscillation state as shown in Fig. 5 will occur, resulting in loss of control. In the case of the present invention, K, , , 1
When KF > 1, a step response as shown in Fig. 9 is obtained, and it is understood that the damping performance is improved by quickly attenuating minute vibrations.Furthermore, when KF > 1, a step response as shown in Fig. 12 is obtained. It can be seen that a response is obtained and the damping properties are extremely improved.

[Brief explanation of drawings]

Fig. 1 is a side view of a painting robot as an example of a direct teaching robot to which the present invention can be applied, and Fig. 2 is a sectional side view of a drive mechanism that can be used in the robot, including a reducer and a clutch. , the installation of the first and second position detectors etc. clearly indicates the state, and Figure 3 shows the steps by the control circuit shown in Figure 3 using only the position detectors provided on the conventional robot element side. An example of a response, FIG. 6 is a block diagram of the entire control device which is an embodiment of the present invention, FIGS. 7 and 8 are block diagrams of the same control device when it is mathematically modeled, and FIG. 9 is a block diagram of the same control device. A graph showing the step response when KF=1 in the control device,
10 and 11 are block diagrams showing the case without position feedback in FIG. 7, and FIG. 12 is a graph showing the step response of the control system shown in FIG. FIG. 13 is a block diagram of the entire control device according to another embodiment. M...1 driving source, k...transmission means 81, 4
9...1st. Second position detection means 106, 202.
...Conversion means, 206...Position storage means 310, 3
16...Gravity balance device. Applicant Kobe Steel Co., Ltd. Agent Patent Attorney Takeshi Honjo Figure 3 i'i', 5 l゛no1 1234 Time t (s) Figure 9 01234 Unexamined Patent Publication t (S) No. 11141 i'1s12r・1 Leg distance t (S)

Claims (1)

[Claims] 1. In a control device for a direct teaching type industrial robot in which a drive source drives a robot element such as an arm via a transmission means such as a reducer, there is provided a first control device for detecting the position of the drive source. position detection means, ■ second position detection means that detects the position of the robot element and feeds back the position information to the input side of the control system, and ■ position detection means obtained by the first and second position detection means. Information power χ et al. Conversion means that calculates position difference information corresponding to the positional deviation between the drive source and the robot element and feeds it back to the input side of the control system, and ■ Sends out from the second position detection means at the time of teaching. a position storage means for sequentially storing teaching position data of the robot element to be taught, ■ a reset means for clearing the difference in position between the first and second position detection means that occurs during teaching, etc. at the time of reproduction, and ■ reproduction. Sometimes the robots listed in ■)! A control means for synthesizing the raw position information and the position difference information described in (■), multiplying the signal of the difference between the synthesized value and the taught position data by a constant gain, and outputting the signal to the drive system. A direct teaching type industrial robot featuring: 2. In a control device for a direct teaching industrial robot in which a robot element such as an arm is driven by a drive source via a transmission means such as a speed reducer, (1) a first position detection means for detecting the position of the drive source; ■A second position detection means that detects the position of the robot element and feeds back the position information to the input side of the control system; ■A drive source is detected from the position information obtained by the first and second position detection means. Teaching device elements consisting of a conversion means that calculates position difference information corresponding to the positional deviation between robot elements and feeds it back to the input side of the control system, and a second position detector during teaching. a position storage means for sequentially storing position data; (1) a reset method for clearing the difference in position between the first and second position detectors that occurs during teaching, etc. during playback; The location information of the robot i described in ■ and ■
a control means that synthesizes the position difference information described in , multiplies the signal of the difference between this composite value and the taught position data by a constant gain, and outputs the signal to the drive system; A control device for an industrial robot, characterized in that it has a gravity balance device (d).